From Volker.Tresp at mchp.siemens.de Fri Oct 1 09:55:33 1999 From: Volker.Tresp at mchp.siemens.de (Volker Tresp) Date: Fri, 01 Oct 1999 15:55:33 +0200 Subject: Two papers on probabilistic approximations to Bayesian networks Message-ID: <37F4BD55.E7D5DDCF@mchp.siemens.de> We would like to announce the availability of two papers. In the first paper we apply different mixture approximations to Bayesian networks with the goal of obtaining insight into the modeled domain. One of the approximations is obtained via a mixture of mean field solutions. The latter approach and a general formulation of the mean field approximation applicable to an arbitrary Bayesian network can be found in the second paper. ------------------------------------------------------------------------- Mixture Approximations to Bayesian Networks Volker Tresp, Michael Haft and Reimar Hofmann Siemens AG, Corporate Technology Neural Computation Dept. Information and Communications Otto-Hahn-Ring 6, 81730 Munich, Germany Published in Laskey, K. B., Prade, H., (Hrsg.), Uncertainty in Artificial Intelligence, Proceedings of the Fifteenth Conference, Morgan Kaufmann Publishers, 1999, pp. 639-646 Structure and parameters in a Bayesian network uniquely specify the probability distribution of the modeled domain. The locality of both structure and probabilistic information are the great benefits of Bayesian networks and require the modeler to only specify local information. On the other hand this locality of information might prevent the modeler ---and even more any other person--- from obtaining a general overview of the important relationships within the domain. The goal of the work presented in this paper is to provide an ``alternative'' view on the knowledge encoded in a Bayesian network which might sometimes be very helpful for providing insights into the underlying domain. The basic idea is to calculate a mixture approximation to the probability distribution represented by the Bayesian network. The mixture component densities can be thought of as representing typical scenarios implied by the Bayesian model, providing intuition about the basic relationships. As an additional benefit, performing inference in the approximate model is very simple and intuitive and can provide additional insights. The computational complexity for the calculation of the mixture approximations critically depends on the measure which defines the distance between the probability distribution represented by the Bayesian network and the approximate distribution. Both the KL-divergence and the backward KL-divergence lead to inefficient algorithms. Incidentally, the latter is used in recent work on mixtures of mean field solutions to which the work presented here is closely related. We show, however, that using a mean squared error cost function leads to update equations which can be solved using the junction tree algorithm. We conclude that the mean squared error cost function can be used for Bayesian networks in which inference based on the junction tree is tractable. For large networks, however, one may have to rely on mean field approximations. http://www7.informatik.tu-muenchen.de/~hofmannr/uai99_abstr.html ----------------------------------------------------------------------- ----------------------------------------------------------------------- Model-Independent Mean Field Theory as a Local Method for Approximate Propagation of Information M.~Haft, R.~Hofmann and V.~Tresp Corporate Technology, Department: Information and Communications Siemens AG, 81730 M\"unchen, Germany Published in Network: Computation in Neural Systems, 10, 1999, pp. 93-105 (based on a TR of 1997). We present a systematic approach to mean field theory (MFT) in a general probabilistic setting without assuming a particular model. The mean field equations derived here may serve as a local and thus very simple method for approximate inference in probabilistic models such as Boltzmann machines or Bayesian networks. Our approach is `model-independent' in the sense that we do not assume a particular type of dependencies; in a Bayesian network, for example, we allow arbitrary tables to specify conditional dependencies. In general, there are multiple solutions to the mean field equations. We show that improved estimates can be obtained by forming a weighted mixture of the multiple mean field solutions. Simple approximate expressions for the mixture weights are given. The general formalism derived so far is evaluated for the special case of Bayesian networks. The benefits of taking into account multiple solutions are demonstrated by using MFT for inference in a small and in a very large Bayesian network. The results are compared to the exact results. http://www7.informatik.tu-muenchen.de/~hofmannr/network99_abstr.html ----------------------------------------------------------------------- From terry at salk.edu Sun Oct 3 23:48:17 1999 From: terry at salk.edu (Terry Sejnowski) Date: Sun, 3 Oct 1999 20:48:17 -0700 (PDT) Subject: NEURAL COMPUTATION 11:8 Message-ID: <199910040348.UAA07616@dax.salk.edu> *** Volume 12 of Neural Computation will comprise 12 monthly issue *** Neural Computation - Contents - Volume 11, Number 8 - November 15, 1999 ARTICLES Detecting And Estimating Signals In Noisy Cable Structure: I. Neuronal Noise Sources Amit Manwani and Christof Koch Detecting And Estimating Signals In Noisy Cable Structures: II. Information-Theoretic Analysis Amit Manwani and Christof Koch Population Dynamics of Spiking Neurons: Fast Transients, Asynchronous States, and Locking Wulfram Gerstner Correctness of Local Probability in Graphical Models with Loops Yair Weiss NOTES Natural Gradient Learning For Over-and-Under-Complete Bases In ICA Shun-ichi Amari Combined 5x2cv F Test For Comparing Supervised Classification Learning Algorithms Ethem Alpaydin LETTER Adaptive Neural Coding Dependent on the Time-varying of the Somatic Input Current Jonghan Shin, Christof Koch, and Rodney Douglas A Reinforcement Learning Approach To On-Line Clustering Aristidis Likas Replicator Equations, Maximal Cliques, And Graph Isomorphism Marcello Pelillo Independent Component Analysis: a Flexible Non-Linearity and Decorrelating Manifold Approach Richard Everson and Stephen Roberts On The Design Of BSB Neural Associative Memories Using Semidefinite Programming Jooyoung Park, Hyuk Cho and Daihee Park How To Design A Connectionist Holistic Parser Ho K. S. Edward and Chan L. Wan A Unified Analysis Of Value-Function-Based Reinforcement- Learning Algorithms Csaba Szepesvari, and Michael Littman Neuronal Regulation: A Mechanism For Synaptic Pruning During Brain Maturation Gal Chechik, Isaac Meilijson and Eytan Ruppin Comparison of SOM Point Densities Based on Different Criteria Teuvo Kohonen ----- NOTE: Neural Computation is now on-line and issues starting with 11:1 are available to all free for a trial period: ON-LINE - http://neco.mitpress.org/ http://intl-neco.mitpress.org (Europe and Asia) ABSTRACT http://intl-neco.mitpress.orgECO/ SUBSCRIPTIONS - 1999 - VOLUME 11 - 8 ISSUES USA Canada* Other Countries Student/Retired $50 $53.50 $84 Individual $82 $87.74 $116 Institution $302 $323.14 $336 * includes 7% GST (Back issues from Volumes 1-10 are regularly available for $28 each to institutions and $14 each for individuals. (Back issues from Volumes 1-10 are regularly available for $28 each to institutions and $14 each for individuals. Add $5 for postage per issue outside USA and Canada. Add +7% GST for Canada.) MIT Press Journals, 5 Cambridge Center, Cambridge, MA 02142-9902. Tel: (617) 253-2889 FAX: (617) 258-6779 mitpress-orders at mit.edu ----- From dimi at ci.tuwien.ac.at Mon Oct 4 10:06:00 1999 From: dimi at ci.tuwien.ac.at (Evgenia Dimitriadou) Date: Mon, 4 Oct 1999 16:06:00 +0200 (CEST) Subject: CI BibTeX Collection -- Update Message-ID: The following volumes have been added to the collection of BibTeX files maintained by the Vienna Center for Computational Intelligence: IEEE Transactions on Evolutionary Computation, Volumes 2/3-3/3 IEEE Transactions on Neural Networks, Volumes 9/3-10/5 Machine Learning, Volumes 31-37 Neural Computation, Volumes 11/4-11/6 Neural Networks, Volumes 12/3-12/8 Neural Processing Letters, Volumes 9/3-10/1 Most files have been converted automatically from various source formats, please report any bugs you find. The complete collection can be downloaded from http://www.ci.tuwien.ac.at/docs/ci/bibtex_collection.html ftp://ftp.ci.tuwien.ac.at/pub/texmf/bibtex/ Best, Vivi ************************************************************************ * Evgenia Dimitriadou * ************************************************************************ * Institut für Statistik * Tel: (+43 1) 58801 10773 * * Technische Universität Wien * Fax: (+43 1) 58801 10798 * * Wiedner Hauptstr. 8-10/1071 * Evgenia.Dimitriadou at ci.tuwien.ac.at * * A-1040 Wien, Austria * http://www.ci.tuwien.ac.at/~dimi* ************************************************************************ From evansdj at aston.ac.uk Mon Oct 4 07:27:35 1999 From: evansdj at aston.ac.uk (DJ EVANS) Date: Mon, 04 Oct 1999 12:27:35 +0100 Subject: RESEARCH FELLOWSHIP: Pollution-in-Water: Detection, tracking and classification Message-ID: <37F88F27.871A8700@aston.ac.uk> Dear All, Please reply to Prof D. Lowe (d.lowe at aston.ac.uk). Thanks, David Evans. ******************************************************************* Neural Computing Research Group ------------------------------- School of Engineering and Applied Sciences Aston University, Birmingham, UK RESEARCH FELLOWSHIP ------------------- Pollution-in-Water: Detection, tracking and classification ---------------------------------------------------------- The Neural Computing Research Group at Aston is looking for a highly motivated individual for a 30 month postdoctoral research position in the area of automatic video analysis to detect, classify and track near-shore sea-borne pollutants. This project will be part of a Framework V European Union collaborative project on an environmental early warning monitoring initiative. Potential candidates should have strong statistical and computational skills, preferably with a background related to neural networks and image analysis. Further information on the project can be obtained from http://www.ncrg.aston.ac.uk/ Conditions of Service --------------------- Salaries will be up to point 9 on the RA 1A scale. The salary scale is subject to review and annual increments. How to Apply ------------ If you wish to be considered for this Fellowship, please send a full CV and publications list, including full details and grades of academic qualifications, together with the names of 3 referees, to: Prof David Lowe Neural Computing Research Group School of Engineering and Applied Sciences Aston University Birmingham B4 7ET, U.K. Tel: 0121 333 4631 Fax: 0121 333 4586 e-mail: d.lowe at aston.ac.uk e-mail submission of postscript/latex files is welcome. Closing date: 13 December, 1999. From ken at phy.ucsf.EDU Tue Oct 5 06:46:20 1999 From: ken at phy.ucsf.EDU (Ken Miller) Date: Tue, 5 Oct 1999 03:46:20 -0700 (PDT) Subject: UCSF Postdoctoral/Graduate Fellowships in Theoretical Neurobiology Message-ID: <14329.55036.61411.427912@coltrane.ucsf.edu> FULL INFO: http://www.sloan.ucsf.edu/sloan/sloan-info.html PLEASE DO NOT USE 'REPLY'; FOR MORE INFO USE ABOVE WEB SITE OR CONTACT ADDRESSES GIVEN BELOW The Sloan Center for Theoretical Neurobiology at UCSF solicits applications for pre- and post-doctoral fellowships, with the goal of bringing theoretical approaches to bear on neuroscience. Applicants should have a strong background and education in mathematics, theoretical or experimental physics, or computer science, and commitment to a future research career in neuroscience. Prior biological or neuroscience training is not required. The Sloan Center offers opportunities to combine theoretical and experimental approaches to understanding the operation of the intact brain. Young scientists with strong theoretical backgrounds will receive scientific training in experimental approaches to understanding the operation of the intact brain. They will learn to integrate their theoretical abilities with these experimental approaches to form a mature research program in integrative neuroscience. The research undertaken by the trainees may be theoretical, experimental, or a combination. Resident Faculty of the Sloan Center and their research interests include: Allison Doupe: Development of song recognition and production in songbirds Stephen Lisberger: Learning and memory in a simple motor reflex, the vestibulo-ocular reflex, and visual guidance of smooth pursuit eye movements by the cerebral cortex Michael Merzenich: Experience-dependent plasticity underlying learning in the adult cerebral cortex, and the neurological bases of learning disabilities in children Kenneth Miller: Circuitry of the cerebral cortex: its structure, self-organization, and computational function (primarily using cat primary visual cortex as a model system) Roger Nicoll: Synaptic and cellular mechanisms of learning and memory in the hippocampus Christoph Schreiner: Cortical mechanisms of perception of complex sounds such as speech in adults, and plasticity of speech recognition in children and adults Michael Stryker: Mechanisms that guide development of the visual cortex There are also a number of visiting faculty, including Larry Abbott, Brandeis University; Bill Bialek, NEC Research; Sebastian Seung, MIT; David Sparks, Baylor University; Steve Zucker, Yale University. TO APPLY, please send a curriculum vitae, a statement of previous research and research goals, up to three relevant publications, and have two letters of recommendation sent to us. The application deadline is February 1, 2000. Send applications to: Steve Lisberger Sloan Center for Theoretical Neurobiology at UCSF Department of Physiology University of California 513 Parnassus Ave. San Francisco, CA 94143-0444 PRE-DOCTORAL applicants with strong theoretical training may seek admission into the UCSF Neuroscience Graduate Program as a first-year student. Applicants seeking such admission must apply by Jan. 5, 2000 to be considered for fall, 2000 admission. Application materials for the UCSF Neuroscience Program may be obtained from http://www.neuroscience.ucsf.edu/neuroscience/admission.html or from Cindy Kelly Neuroscience Graduate Program Department of Physiology University of California San Francisco San Francisco, CA 94143-0444 neuroscience at phy.ucsf.edu Be sure to include your surface-mail address. The procedure is: make a normal application to the UCSF Neuroscience program; but also alert the Sloan Center of your application, by writing to Steve Lisberger at the address given above. If you need more information: -- Consult the Sloan Center WWW Home Page: http://www.sloan.ucsf.edu/sloan -- Send e-mail to sloan-info at phy.ucsf.edu -- See also the home page for the W.M. Keck Foundation Center for Integrative Neuroscience, in which the Sloan Center is housed: http://www.keck.ucsf.edu/ From danr at cs.uiuc.edu Wed Oct 6 10:01:20 1999 From: danr at cs.uiuc.edu (Dan Roth) Date: Wed, 06 Oct 1999 09:01:20 -0500 Subject: Software release: SNoW Learning Architecture References: <199910051432.PAA00954@tahiti> Message-ID: <37FB5630.5AEC22AA@cs.uiuc.edu> ---------------------------------------------------------------------- Software release: SNoW Learning Architecture 2.0 Cognitive Computation Group University of Illinois at Urbana/Champaign http://L2R.cs.uiuc.edu/~cogcomp/ ---------------------------------------------------------------------- The Cognitive Computation Group at the University of Illinois at Urbana/Champaign announces the release of the SNoW Learning Architecture (version 2.0). SNoW is a learning program that can be used as a general purpose multi-class classifier and is specifically tailored for learning in the presence of a very large number of features. The learning architecture is a sparse network of linear units over a pre-defined or incrementally acquired feature space. Several update rules may be used - sparse variations of the Winnow update rule, Perceptron, or naive Bayes. SNoW has been used successfully in several applications in the natural language and visual processing domains; the release is meant to be used only for research purposes, with the hope that it can be a useful research tool for studying learning in these domains. Feedback of any sort is welcome. You are invited to download the SNoW package for educational or non-commercial research purposes. When downloading the package you are asked to register and express your agreement with the license terms, under the university of Illinois guidelines. SNoW is not shareware or public domain software. The SNoW software package can be downloaded by following the `Software' link under the Cognitive Computation group home page at http://L2R.cs.uiuc.edu/~cogcomp The SNoW package contains the following: - Source code (C++) with a Makefile. - A user guide (UIUC Tech Report, UIUC-DCS-R-99-210) containing brief description of the architecture and algorithms, detailed descriptions of the command-line options, and a brief tutorial. - The text of the license agreement. Two packages are available, one for UNIX system (that should be easily installed also on most Linux systems) and a second for NT systems. The User guide (UIUC Tech report UIUC-DCS-R-99-210, by Andrew J. Carlson, Chad M. Cumby, Jeff L. Rosen and Dan Roth) can be downloaded directly from: http://L2R.cs.uiuc.edu/~danr/Papers/userguide.ps.gz Papers related to SNoW can be found at http://L2R.cs.uiuc.edu/~danr/snow.html For comments and bug reports relating to SNoW, please send mail to SNoW at cs.uiuc.edu ---------------------------------------------------------------------- Dan Roth Department of Computer Science, University of Illinois, Urbana/Champaign 1304 W. Springfield Ave. Urbana IL 61801 Phone: (217) 244-7068 (217) 244-6813 (Sec) Fax: +(217) 244-6500 e-mail: danr at cs.uiuc.edu http://L2R.cs.uiuc.edu/~danr ---------------------------------------------------------------------- From steve at cns.bu.edu Wed Oct 6 20:25:03 1999 From: steve at cns.bu.edu (Stephen Grossberg) Date: Wed, 6 Oct 1999 20:25:03 -0400 Subject: How The Basal Ganglia Learn to Selectively Respond to Unexpected Rewarding Cues Message-ID: The following article can be read at http://cns-web/bu.edu/Profiles/Grossberg HOW THE BASAL GANGLIA USE PARALLEL EXCITATORY AND INHIBITORY LEARNING PATHWAYS TO SELECTIVELY RESPOND TO UNEXPECTED REWARDING CUES. Brown, J., Bullock, D., and Grossberg, S. (1999). Journal of Neuroscience, in press. After classically conditioned learning, dopaminergic cells in the substantia nigra pars compacta (SNc) respond immediately to unexpected conditioned stimuli (CS) but omit formerly seen responses to expected unconditioned stimuli, notably rewards. These cells play an important role in reinforcement learning. A neural model explains the key neurophysiological properties of these cells before, during, and after conditioning, as well as related anatomical and neurophysiological data about the pedunculo-pontine tegmental nucleus (PPTN), lateral hypothalamus, ventral striatum, and striosomes. The model proposes how two parallel learning pathways from limbic cortex to the SNc, one devoted to excitatory conditioning (through the ventral striatum, ventral pallidum, and PPTN) and the other to adaptively timed inhibitory conditioning (through the striosomes), control SNc responses. The excitatory pathway generates CS-induced excitatory SNc dopamine bursts. The inhibitory pathway prevents dopamine bursts in response to predictable reward-related signals. When expected rewards are not received, striosomal inhibition of SNc that is unopposed by excitation results in a phasic drop in dopamine cell activity. The adaptively timed inhibitory learning uses an intracellular spectrum of timed responses that is proposed to be similar to adaptively timed cellular mechanisms in the hippocampus and the cerebellum. These mechanisms are proposed to include metabotropic glutamate receptor-mediated Ca2+ spikes that occur with different delays in striosomal cells. A dopaminergic burst in concert with a Ca2+ spike is proposed to potentiate inhibitory learning. The model provides a biologically predictive alternative to temporal difference (TD) conditioning models and explains substantially more data than alternative models. Keywords: dopamine, substantia nigra, reward, basal ganglia, conditioning, pedunculopontine tegmental nucleus, lateral hypothalamus, striosomes, adaptive timing Preliminary version appears as Boston University Technical Report, CAS/CNS-TR-99-011. Download this paper as: Gzipped Postscript BroBulGro99.ps.gz (152Kb) or PDF (BroBulGro99.pdf) From gary at cs.ucsd.edu Wed Oct 6 21:42:52 1999 From: gary at cs.ucsd.edu (Gary Cottrell) Date: Wed, 6 Oct 1999 18:42:52 -0700 (PDT) Subject: Jobs at UCSD Message-ID: <199910070142.SAA20655@gremlin.ucsd.edu> Hi folks - This is a "heads up" that the CSE Department at UCSD has several faculty positions at Junior AND Senior levels that we are going to try to fill this coming year. While the ad specifies many areas, among the ones listed are machine learning, data mining, bioinformatics, and computational biology. The goal is to hire the best candidates, regardless of area. If you are one of those, please apply! See the web page: http://www-cse.ucsd.edu/cse/Academic.Jobs.html The deadline is January 15, 2000. The CSE Department at UCSD is consistently ranked among the top 20 in the US, and has made several key hires in recent years in theory, networking, distributed systems. UCSD has been ranked the number 1 public science university in the US using only objective measures. There is a broad group of people around the university interested in various standard combinations of the words Neural, Computation, Cognition, and Science. Plus, we have the best weather - the beach party last year was on Halloween! cheers, gary Gary Cottrell 858-534-6640 FAX: 858-534-7029 Faculty Assistant Chet Frost: 858-822-3286 Computer Science and Engineering 0114 IF USING FED EX INCLUDE THE FOLLOWING LINE: "Only connect" 3101 Applied Physics and Math Building University of California San Diego -E.M. Forster La Jolla, Ca. 92093-0114 Email: gary at cs.ucsd.edu or gcottrell at ucsd.edu Home page: http://www-cse.ucsd.edu/~gary/ From D.Willshaw at cns.ed.ac.uk Thu Oct 7 06:46:55 1999 From: D.Willshaw at cns.ed.ac.uk (David Willshaw) Date: Thu, 7 Oct 1999 11:46:55 +0100 (BST) Subject: NETWORK: Computation in Neural Systems Message-ID: <14332.31263.408603.380963@gargle.gargle.HOWL> `NETWORK: Computation in Neural Systems' has changed its scope. A shortened version of the Editorial which will appear in the next issue follows. CHANGES AT `NETWORK: COMPUTATION IN NEURAL SYSTEMS' Journals such as `NETWORK: Computation in Neural Systems' have exploited the essential ambiguity in such terms as `computation in neural systems' by appealing both to those interested in the nervous system and those interested in neurally-inspired artefacts and algorithms. However, with the rapid expansion seen in both areas of research, there would come a time when it would be practicable for the journal to focus onto one of these research areas. This time has come. Given the underlying interest within `NETWORK' in the nervous system, it is natural that the journal now focusses on `computational neuroscience'. The journal provides a forum for integrating theoretical and experimental findings in computational neuroscience across relevant interdisciplinary boundaries. It aims to make theor- etical results and methods accessible to neurobiologists, psychologists and cognitive scientists. Similarly, the rapidly accumulating empirical data in the neurobiological, psych- ological and cognitive domains will enable theorists to stimulate a synthesis or provoke new models. This is the new scope of `NETWORK: Computation in Neural Systems' which is published in every issue. `NETWORK' spans the disciplines of mathematics, physics, computer science, psychology, cognitive science, medicine, neurobiology, amongst others. Work on `computation in neural systems' as defined here refers to work on theoretical and computational aspects of the development and functioning of the nervous system, which can be at the level of networks of neurons or at the cellular or the subcellular level; or reporting new experimental neuroscience findings which are presented within the context of a theory or model; or reporting the application of more abstract, neural network models to problems at the psychological, cognitive and linguistic levels; or developing and analysing novel theories and models which have a strong and direct inspiration from natural systems such as neurobiological systems. In practical terms, the only change discernable in the contents of the journal will be that work on neural networks without a strong biological application or inspiration other than that expressed in the term `neural networks' now falls outside the scope of the journal. Recent Organisational Changes at `NETWORK' Along with the change in scope there has been an organisation change. The journal's Editorial office has moved to Edinburgh from the publishers, Institute of Physics Publishing (IOPP), in Bristol, UK. Following the spirit of the closer Editorial involvement with the journal that this change implies, members of the Editorial Board and I welcome at all times comments and suggestions from authors, readers and referees that will help us to improve the journal according to the needs of the community we serve. Abstracting and Indexing `NETWORK is covered by the leading indexing services including Science Citation Index, Neuroscience Citation Index, Current Contents and Index Medicus/MEDLINE. `NETWORK' and Electronic Publishing IOPP has extensive experience in electronic publishing. Here are some of the current and future benefits of electronic publishing for authors and readers of `NETWORK': Papers that are submitted electronically can appear on the website very soon after they have been accepted, which can be at least a month before the printed journal appears. Electronic versions of papers can include full colour figures, and multimedia attachments can be provided free of charge; see www.iop.org/Journals/nfa for further details. IOPP Electronic Journals were accessed over 2.5 million times in 1998, of which 22,000 hits were aimed at `NETWORK' articles. As from 1st January 2000, readers of all IOPP journals will have access to the last 10 years on-line; this means that the whole of `NETWORK' will be accessible in this way. In the longer term, IOPP is developing a web-based system of refereeing that should lead to even faster publication times. In Conclusion For full information on the changes at `NETWORK', an Editorial will feature in the next issue of the journal. We are confident that the recent changes can only benefit authors and readers of `NETWORK. I am looking forward to establishing the journal as a premier European and world-wide contributor to the field of computational neuroscience. Professor David Willshaw Editor-in-Chief, `NETWORK: Computation in Neural Systems' Institute for Adaptive and Neural Computation Division of Informatics University of Edinburgh 2 Buccleuch Place Edinburgh EH8 9LW Scotland, UK Tel: +44 (0)131 650 4404/5 Fax: +44 (0)131 650 4406 Email: neted at anc.ed.ac.uk From cmbishop at microsoft.com Fri Oct 8 10:14:33 1999 From: cmbishop at microsoft.com (Christopher Bishop) Date: Fri, 8 Oct 1999 07:14:33 -0700 Subject: Postdoctoral Research Fellowship at Cambridge Message-ID: <3FF8121C9B6DD111812100805F31FC0D101F2650@RED-MSG-59> DARWIN COLLEGE, CAMBRIDGE Microsoft Research Fellowship http://www.dar.cam.ac.uk/ http://research.microsoft.com/cambridge Closing date: 15 October 1999. The Governing Body of Darwin College Cambridge, and Microsoft Research Limited, jointly invite applications for a stipendiary Research Fellowship supporting research in the field of adaptive computing (including topics such as pattern recognition, probabilistic inference, statistical learning theory and computer vision). Eligibility Men and women graduates of any university are eligible to apply, irrespective of age, provided they have a doctorate or an equivalent qualification, or expect to have submitted their thesis before taking up the Fellowship. Tenure The Fellowship will be tenable for two years commencing l October 2000 or on a date to be agreed. The successful candidate will engage in research full-time at the Microsoft Research Laboratory in Cambridge, and will join the new Machine Learning and Perception research group currently comprising Chris Bishop, Andrew Blake, Ben Bradshaw, Bernhard Schoelkopf, Michael Tipping and Phil Torr. The Fellow will be a member of the Governing Body of Darwin College and will be subject to the Statutes and Ordinances of the College which may be seen on request to the Bursar. The Statutes include the obligation to reside in or near Cambridge for at least two-thirds of each University term, but the Governing Body will normally excuse absences made necessary by the nature of the research undertaken. Stipend and Emoluments The stipend will be dependent upon age and experience. In addition the Fellow will be able to take seven meals per week at the College table free of charge and additional meals at his or her own expense. Guests may be invited to all meals (within the limits of available accommodation), ten of them free of charge within any quarter of the year. College accommodation will be provided, subject to availability, or an accommodation allowance will be paid in lieu. In addition to a salary the Fellowship provides funding for conference participation. Applications Applications should reach the Master, Darwin College, Cambridge CB3 9EU by 15 October 1999. They should be typed and should include SIX copies of (1) a curriculum vitae, (2) an account, in not more than 1000 words, of the proposed research, including a brief statement of the aims and background to it, (3) the names and addresses of three referees (including telephone, fax and e-mail co-ordinates), WHO SHOULD BE ASKED TO WRITE AT ONCE DIRECT TO THE COLLEGE, and (4) a list of published or unpublished work that would be available for submission if requested. Testimonials should not be sent. Electronic applications (in postscript, pdf, Word or text formats) may be sent to cmbishop at microsoft.com to arrive by 15 October 1999. Short-listed candidates may be asked to make themselves available for interview at Darwin College on a date to be arranged before mid-January 2000: election will be made as soon as possible thereafter. In certain circumstances travelling expenses for overseas interviewees may be covered. The College follows an equal opportunities policy. From dagli at umr.edu Sat Oct 9 15:31:33 1999 From: dagli at umr.edu (Cihan Dagli) Date: Sat, 9 Oct 1999 14:31:33 -0500 Subject: Artificial Neural Networks in Engineering Conference ANNIE'99 Message-ID: Dear Colleagues On behalf of the organizing committee I would like to invite you to attend ANNIE 1999, an international conference to be held on November 7-10, 1999, at Marriott's Pavilion Hotel in downtown, St. Louis, Missouri, USA. This will be the ninth international gathering of researchers interested in Smart Engineering System Design using neural networks, fuzzy logic, evolutionary programming, data mining, and complex systems. The previous conferences each drew approximately 150 papers from twenty countries with their proceedings published by ASME Press as hardbound books in eight volumes. The last volume, edited by Dagli, et. al., was titled ``Smart Engineering Systems: Neural Networks, Fuzzy Logic, Evolutionary Programming, Data Mining, and Rough Sets.'' Enclosed Please find the preliminary program for the conference. You can get the latest information about the conference at http://www.umr.edu/~annie/annie99/ and for the details of advanced registration to the conference at http://www.umr.edu/~annie/rf.htm . The conference will cover the theory of Smart Engineering System Design techniques, namely; neural networks, fuzzy logic, evolutionary programming, data mining, and complex systems. Presentations dealing with applications of these technologies are encouraged in the areas of: manufacturing engineering, biology and medicine, pattern recognition, image processing, process monitoring, control, recent theoretical developments in neural networks, fuzzy logic, data mining, rough sets, evolutionary programming, fractals, chaos, and wavelets that can impact Smart Engineering System Design. The response to ANNIE '99 was excellent with 310 abstracts received. Papers submitted based on these abstracts were reviewed by two referees and all accepted papers are included in the conference proceedings to be published by ASME Press as a hardbound book titled Smart Engineering System Design: Neural Networks, Fuzzy Logic, Evolutionary Programming, Data Mining and Complex Systems, edited by Drs. Dagli, Buczak, Ghosh, Embrechts and Ersoy. Conference banquet and the presentation of Best Paper awards are scheduled on Tuesday evening. Our Banquet Plenary Speaker for this year's ANNIE is University Professor Simon Haykin. He is the director of Neurocomputation for Signal Processing Group at McMaster University, Hamilton, Ontario, Canada. His banquet talk is titled "Neural Networks for Signal Processing." There are nine plenary sessions scheduled for ANNIE '99. Dr. Walter Freeman will present "The Neurodynamics of Intentionality is the Basis of Intelligent Behavior" at the Monday morning plenary session. At noon on Monday, Dr. Hojjat Adeli will present "Distributed and Neurocomputing for Large-Scale Engineering Design Automation." Monday afternoon Dr. Bhavani Thuraisingham will present " Data Mining and Knowledge Discovery: Developments and Challenges". In Tuesday plenary session, " Evolvable Hardware" will be presented by Dr. Adrian Stoica. Tuesday's luncheon plenary "Intangled Behavior and Bizarre Systems" will be presented by Dr. Steve Kercel. The Tuesday afternoon plenary will be "Neural Mechanics of the Cerebellum and Its Application" presented by Dr. Witali Dunin-Barkowski. Dr. Kristin Bennett will present "Support Vector Machines: An Overview and New Extensions" for the Wednesday morning plenary. The luncheon plenary " Searching the Web-it's worse than you think!" will be presented by Dr. Lee Giles. The Wednesday afternoon plenary will be "Neural Networks for Analysis of Data with Gaps" presented by Dr. Alexander Gorban. Half-day tutorials have been scheduled for Sunday, November 7th. These state-of-the-art workshops cover the following areas: Collective Intelligence Chaos Data Mining: Technologies, Tools and Trends Hybrid Intelligent Systems Using Soft Computing Techniques and Fractal Theory Computer Vision and Fuzzy-Neural Network Models Classifiers Ensembles: how and why they work Interior Point Optimization Methods in Support Vector Machines Training Novel Applications of Neural Network, Signal Processing and Data Mining Technologies An Introduction to Evolutionary Engineering Pattern Recognition Applications Development: The Process I would like to thank members of the Organizing Committee, Sponsoring Organization, and Co-Chairs, Drs. Buczak, Ghosh, Embrechts and Ersoy, for putting together an excellent program for ANNIE '99. I would like to recognize the excellent and timely efforts of the referees and contributors that made the conference possible. We are looking forward to your visit and participation at the meeting. Sincerely, Cihan H. Dagli Conference Chairman Smart Engineering Systems Laboratory 229 Engineering Management 1870 Miner Circle University of Missouri-Rolla Rolla MO 65409-0370 Phone: 573 341-4374 Fax: 573 341-6567 USA http://www.umr.edu/~annie http://www.gbhap.com/Smart_Engineering_System_Design/ http://www.umr.edu/~dagli From cg at santafe.edu Fri Oct 8 18:23:28 1999 From: cg at santafe.edu (Christine Gonzales) Date: Fri, 8 Oct 1999 16:23:28 -0600 Subject: Complex Systems Summer School, 2000 Message-ID: Santa Fe Institute Complex Systems Summer School, 2000 June 4 to June 30, 2000 in Santa Fe, New Mexico USA An intensive four-week introduction to complex behavior in mathematical, physical and biological systems, intended for graduate students and postdoctoral fellows. No tuition is charged. The first week will consist of an intensive series of lectures, demonstrations, and experiments introducing some core ideas and tools of complex systems research. The topics will include experimental and theoretical nonlinear dynamics and pattern formation, statistical mechanics and stochastic processes, information theory and computation theory, adaptive computation, and computer modeling tools. (Prior to week one, there may be an optional intensive two-day math review.) Weeks two and three will consist of short courses on current research in complex systems, and students will work on individual or team projects. Week four will be devoted to completion and presentation of student projects. Co-Directors: Raymond E. Goldstein, Physics and Applied Mathematics, University of Arizona; Melanie Mitchell, Biophysics, Los Alamos National Laboratory; and Lynn Nadel, Psychology, University of Arizona. Location: Held on the campus of the College of Santa Fe. Administered by the Santa Fe Institute. Application instructions: Provide a current resume with publications list (if available), statement of current research interests, comments about why you want to attend the school, and two letters of recommendation from scientists who know your work. Include your e-mail address and fax number. Send only complete application packages by postal mail to: Summer School, Santa Fe Institute 1399 Hyde Park Road Santa Fe, New Mexico USA 87501 505-984-8800 ext. 235 (v); 505-982-0565 (fax) February 7, 2000 deadline. Women and minorities encouraged to apply. Further information at http://www.santafe.edu/sfi/education/indexCSSS.html or summerschool at santafe.edu Christine Gonzales (505) 984-8800 ext. 235 Fax. 982-0565 From vaina at engc.bu.edu Mon Oct 11 15:29:11 1999 From: vaina at engc.bu.edu (Lucia M. Vaina) Date: Mon, 11 Oct 1999 15:29:11 -0400 Subject: Research position in fMRI applied to neurorecovery after stroke Message-ID: A non-text attachment was scrubbed... Name: not available Type: text/enriched Size: 2167 bytes Desc: not available Url : https://mailman.srv.cs.cmu.edu/mailman/private/connectionists/attachments/00000000/d562bf7d/attachment.bin From ingber at ingber.com Mon Oct 11 22:28:30 1999 From: ingber at ingber.com (Lester Ingber) Date: Mon, 11 Oct 1999 21:28:30 -0500 Subject: EEG data available Message-ID: <19991011212830.A27943@ingber.com> I have had several requests for EEG data since I had to remove access from a large dataset I used for some of my studies, due to a disk crash at an archive at U Oregon. I am making a small dataset (4 MB) and two larger datasets (40 MB each) available at my Caltech archive. I do not know for how long I will be able to keep the larger datasets available. I cannot keep the largest dataset (700 MB) available there and do not plan on uploading it anyplace else. All files sizes are in .gz-compressed formats. See http://www.ingber.com/smni_eeg_data.html I hope you find this data useful for your own studies. Lester -- Lester Ingber http://www.ingber.com/ PO Box 06440 Wacker Dr PO Sears Tower Chicago IL 60606-0440 http://www.alumni.caltech.edu/~ingber/ From melnik at data.cs.brandeis.edu Tue Oct 12 12:55:41 1999 From: melnik at data.cs.brandeis.edu (Ofer Melnik) Date: Tue, 12 Oct 1999 12:55:41 -0400 (EDT) Subject: Network analysis page Message-ID: I'ld like to invite you to check out the webpage we put together to illustrate the Decision Intersection Boundary Algorithm (DIBA). The DIBA algorithm is used to analyze feed-forward threshold neural networks, by generating polytopics decision regions that describe network function. It makes a great pedagogical tool, clearly illustrating what neural networks do and how they learn. The page has multiple examples of networks analyzed with the algorithm, MOVIES of networks learning, some theoretical results and a link to the Tech-Report. The URL is: http://www.demo.cs.brandeis.edu/pr/DIBA/ -Ofer ----------------------------------------------------------------- Ofer Melnik melnik at cs.brandeis.edu Volen Center for Complex Systems Ph: (781)-736-2719 Brandeis University (781)-736-DEMO Waltham MA 02254 From pascal at icsc.ab.ca Tue Oct 12 12:27:27 1999 From: pascal at icsc.ab.ca (pascal@icsc.ab.ca) Date: Tue, 12 Oct 1999 10:27:27 -0600 Subject: Neural Computation Symposium Message-ID: <19991012162730Z25908-276+18008@mail.compusmart.ab.ca> ANNOUNCEMENT / CALL FOR PAPERS Second International ICSC Symposium on NEURAL COMPUTATION / NC'2000 To be held at the Technical University of Berlin, Germany May 23-26, 2000 http//www.icsc.ab.ca/nc2000.htm SYMPOSIUM CHAIR Prof. Hans Heinrich Bothe Technical University of Denmark (DTU) Department of Information Technology Building 344 DK-2800 Lyngby, Denmark Email: hhb at it.dtu.dk Phone: +45-4525-3632 Fax: +45-4588-0117 PUBLICATION CHAIR Prof. Raul Rojas Freie Universit?t Berlin Institut Informatik / FB Mathematik Takustrasse 9 D - 14195 Berlin / Germany Email: rojas at inf.fu-berlin.de Fax: +49-30-8387-5109 SYMPOSIUM ORGANIZER ICSC International Computer Science Conventions P.O. Box 279 Millet, Alberta T0C 1Z0 / Canada Phone: +1-780-387-3546 Fax: +1-780-387-4329 Email: operating at icsc.ab.ca WWW: http//www.icsc.ab.ca INTERNATIONAL PROGRAM COMMITTEE Igor Aleksander, Imperial College of Science & Technology, London, U.K. Peter G. Anderson, Rochester Institute of Technology, NY, USA Horst Bischof, Technical University Vienna, Austria Ruediger W. Brause, J.W. Goethe-University, Frankfurt, Germany Juan Lopez Coronado, Universidad Polotecnica de Cartagena, Spain Ludwig Cromme, Brandenburgische Technische Universitaet Cottbus, Germany Chris deSilva, University of Western Australia Crawley, Australia Georg Dorffner, Austrian Research Institute for Artificial Intelligence, Vienna, Austria Gunhan Dundar, Bogazici University, Istanbul, Turkey Bernd Fritzke, Dresden University of Technology, Dresden, Germany Kunihiko Fukushima, The University of Electro-Communications, Tokyo, Japan Wulfram Gerstner, EPFL Lausanne, Switzerland Stan Gielen, University of Nijmegen, Netherlands Marco Gori, University of Siena, Italy Bruce Graham, University of Edinburgh, U.K. Dorothea Heiss-Czedik, Technical University Vienna, Austria Michael Heiss, Siemens Vienna, Austria Giacomo Indiveri, Swiss Federal Institute of Technology, Zurich, Switzerland Lakhmi C. Jain, University of South Australia, The Levels, Australia Nikola Kasabov, University of Otago, Dunedin, New Zealand Bart Kosko, University of Southern California, Los Angeles CA, USA Rudolf Kruse, University of Magedburg, Germany Te-Won Lee, Salk Institute & University of California, San Diego, USA Fa-Long Luo, R&D Department, Redwood City, USA G.Nicolas Marichal, University of La Laguna Tenerife, Spain Giuseppe Martinelli, University of Rome 1, Italy Klaus-Robert Mueller, GMD First, Berlin, Germany Fazel Naghdy, University of Wollongong, Australia M. Palaniswami, University of Melbourne, Australia Guenther Palm, University of Ulm, Germany Yoh-Han Pao, Case Western Reserve University, Cleveland OH, USA Alexander V. Pavlov, Laboratory for Optical Fuzzy Systems, St. Petersburg, Russia Witold Pedrycz, University of Alberta, Edmonton, Canada Raul Rojas, Freie Universit?t Berlin, Germany V. David Sanchez A., Falon, Inc., San Diego CA, USA Bernd Schuermann, Siemens ZFE, Munich, Germany J.S. Shawe-Taylor, Royal Holloway University of London, U.K. Peter Sincak, Technical University of Kosice, Slovakia Nigel Steele, Coventry University, U.K. Rainer Stotzka, Forschungszentrum Karlsruhe, Germany Piotr Szczepaniak, Technical University of Lodz, Poland Csaba Szepesvari, University of Szeged, Hungary Henning Tolle, Technische Hochschule Darmstadt, Germany Shiro Usui, Toyohashi University of Technology, Toyohashi, Japan Luis Alfredo Vidal de Carvalho, Federal University of Rio de Janeiro, Brazil Roberto C. Villas-Boas, United Nations Industrial Development Organization, Rio de Janeiro, Brazil Lipo Wang, Nanyang Technological University, Singapore Klaus Weber, Technical University of Cottbus, Germany Andreas Weingessel, Technical University Vienna, Austria Takeshi Yamakawa, Kyushu Institute of Technology, Fukuoka, Japan Andreas Zell, Universitaet Tuebingen, Germany Tom Ziemke, University of Skoevde, Sweden Jacek M. Zurada, University of Louisville, K.Y., USA ************************************************* SPONSORS Technical University of Berlin, Germany Fraunhofer-Gesellschaft fuer Produktionsanlagen, Berlin, Germany (FHG-IPK) ICSC International Computer Science Conventions, Canada/Switzerland ************************************************* INTRODUCTION The science of neural computation focusses on mathematical aspects to solve complex practical problems, and it also seeks to help neurology, brain theory and cognitive psychology in the understanding of the functioning of the nervous system by means of computational models of neurons, neural nets and subcellular processes. NC'2000 aims to become a major point of contact for research scientists, engineers and practitioners throughout the world in the field of Neural Computation. Participants will share the latest research, developments and ideas in the wide arena of disciplines encompassed under the heading of NC'2000 as a follow-up of the most successful NC'98 conference in Vienna, Austria. ************************************************* TOPICS Contributions are sought in areas based on the list below, which is indicative only. Contributions from new applications areas are welcome. COMPUTATIONAL NEURAL NETWORK MODELS - Artificial neural network paradigms - Knowledge representation - Learing and generalization - Probabilistic neural networks - Information theoretic approaches - Time-coded neural networks - Pulse-coded neural networks - Self-organization - Cellular automata - Hybrid systems (e.g. neuro-fuzzy, GA, evolutionary strategies) - Chaos in neural networks - Statistical and numerical aspects NEUROPHYSIOLOGICALLY INSPRIED MODELS - Neurophysiological foundations - Spiking neuron models and neuron assemblies - Models of brain centers and sensory pathways - Sensormotor integration - Sensation, Perception and Attention - Spatio-temporal Orientation - Reactive Behavior SOFTWARE AND HARDWARE IMPLEMENTATIONS - Simulation and Graphical Programming Tools - Distributed Systems - Neuro-chips, -controllers and -computers - Analog and Digital Electronic Implementations - Optic, Holographic Implementations NEURAL NETWORK APPLICATIONS - Pre-processing and Feature Extraction - Sound, Speech and Image Processing - Pattern Recognition and System Identification - Computer Vision, Feature Binding and Image Understanding - Autonomous Sensor Systems, Multivariate Sensor Fusion - Robotics and Control - Behavior based Exploration and Planning - Power Systems - Environmental Systems - Decision Support Systems - Medical Applications - Operational Research and Logistics ************************************************* SCIENTIFIC PROGRAM NC'2000 will include invited plenary talks, contributed sessions, invited sessions, workhops and tutorials. At present, the following special sessions are in preparation SELF-ORGANIZING MAPS MODIFICATION, EXTENSION AND APPLICATION with a tutorial by Prof. Teuvo Kohonen, who is the 'inventor' of SOM's Session Chair Dr. Udo Seiffert, University of Magdeburg, Germany NONLINEAR DYNAMICS AND NEURAL FIELDS FOR THE ANALYSIS AND ORGANIZATION OF BEHAVIOR with participation by Prof. S.-I. Amari, Riken Institute, Japan, who is the founder of the mathematical formulation of neural fields. Session Chair Prof. Werner von Seelen, Ruhr-University Bochum, Germany ANN IN BIOMEDICINE Session Chair Prof. Piotr Szczepaniak, Technical University of Lodz, Poland OPTICS IN FUZZY SET THEORY Session Chair Dr. Alexander V. Pavlov, Laboratory for Optical Fuzzy Systems, St. Petersburg, Russia ************************************************* INVITED SESSIONS The organization of invited sessions is encouraged. Prospective organizers are requested to send a session proposal (consisting of 4-5 invited papers, the recommended session-chair and co-chair, as well as a short statement describing the title and the purpose of the session to the Symposium Chairman or the Symposium Organizer. Invited sessions should preferably start with a tutorial paper. The registration fee of the session organizer will be waived, if at least 4 authors of invited papers register to the conference. ************************************************* POSTER PRESENTATIONS Poster presentations are encouraged for people who wish to receive peer feedback and practical examples of applied research are particularly welcome. Poster sessions will allow the presentation and discussion of respective papers, which will also be included in the conference proceedings. ************************************************* WORKSHOPS, TUTORIALS AND OTHER CONTRIBUTIONS Proposals should be submitted as soon as possible to the Symposium Chairman or the Symposium Organizer. ************************************************* SUBMISSION OF PAPERS Prospective authors are requested to either send a draft paper (maximum 7 pages) or an extended abstract for review by the International Program Committee. All papers must be written in English, starting with a succinct statement of the problem, the results achieved, their significance and a comparison with previous work. Submissions must be received by October 31, 1999. Regular papers, as well as poster presentations, tutorial papers and invited sessions are encouraged. The submission should also include - Title of conference (NC'2000) - Type of paper (regular, poster, tutorial or invited) - Title of proposed paper - Authors names, affiliations, addresses - Name of author to contact for correspondence - E-mail address and fax # of contact author - Topics which best describe the paper (max. 5 keywords) - Short CV of authors (recommended) Contributions are welcome from those working in industry and having experience in the topics of this conference as well as from academics. The conference language is English. It is strongly recommended to submit abstracts by electronic mail to operating at icsc.ab.ca or else by fax or mail (2 copies) to the following address ICSC Canada P.O. Box 279 Millet, Alberta T0C 1Z0 Canada Fax +1-780-387-4329 ************************************************* BEST PRESENTATION AWARDS The best oral and poster presentations will be honored with best presentation awards. ************************************************* PUBLICATIONS Conference proceedings (including all accepted papers) will be published by ICSC Academic Press and be available for the delegates at the symposium in printed form or on CD-ROM. Authors of a selected number of innovative papers will be invited to submit extended manuscripts for publication in prestigious international journals. ************************************************* IMPORTANT DATES - Submission Deadline: October 31, 1999 - Notification of Acceptance: December 31, 1999 - Delivery of full papers: February 15, 2000 - Tutorials and Workshops: May 23, 2000 - NC'2000 Symposium: May 24-26, 2000 ************************************************* ACCOMMODATION Accommodation at reasonable rates will be available at nearby hotels. Full details will follow with the letters of acceptance. ************************************************* SOCIAL AND TOURIST ACTIVITIES A social program will be organized and also be available for accompanying persons. ************************************************* THE FASCINATING CITY OF BERLIN The old and new capital of Germany is a mecca for scientists and cultural enthusiasts, for day workers and night owls. Charming with its several opera houses, concert halls, cabarets, and beer gardens, Berlin is full of spontaneous cultural events or happenings. As an urban building site for the future, it is at the same time a living contradiction and a casual place with relaxation areas and large parks right in the city center. The fine nature, pine tree forests, and 1001 lakes around the city supply Berlin with it s very specific sparkling air in spring time. No other city in Germany has during the last 100 years played such a prominent role in history and in the imagination of the people Social and industrial revolution in the last century, world war I and later manifestation of the first German republic, the 'Golden Twenties', nazis dictatorship and world war II, splitting by the Berlin Wall, 'economic miracle' in the west and socialistic showpiece city in the east, '68 stu dent and alternative lifestyle movement in the west and peace movement in the east, and finally, the fall of the Wall. Berlin tempts with its many research facilities, and it is Germany's largest industrial city with headquarters or dependences of most major companies. At present, approximately 3.5 million inhabitants live in the reunified city, among which more than 120.000 students, who study in three universities and twelve colleges or schools of arts. Berlin ist eine Reise wert. Welcome! ************************************************* FURTHER INFORMATION Fully updated information is available from http//www.icsc.ab.ca/nc2000.htm You may also contact - ICSC International Computer Science Conventions P.O. Box 279 Millet, Alberta T0C 1Z0 / Canada Phone: +1-780-387-3546 Fax: +1-780-387-4329 Email: operating at icsc.ab.ca or, for specific scientific requests, the symposium chairman. From pcohen at cse.ogi.edu Tue Oct 12 17:44:52 1999 From: pcohen at cse.ogi.edu (Phil Cohen) Date: Tue, 12 Oct 1999 14:44:52 -0700 Subject: Postdoctoral Position: Gesture recognition at OGI Message-ID: <007601bf14fb$03f41d60$78365f81@cse.ogi.edu> The following position is available at the Center for Human-Computer Communication at the Oregon Graduate Institute: TOPIC: Two-dimensional and three-dimensional gesture recognition as part of multimodal systems. You will be an integral member of a multidisciplinary team that has developed leading edge technology. Topics will include gesture interaction for 2-D map-based interactions, and also for 3-D real and virtual worlds. EXPERIENCE: Postdoctoral researcher (new Ph.D. or more experienced), extensive knowledge of neural networks (especially committee-based approaches), and HMMs. Prior work with gesture recognition and/or handwriting recognition is essential. Experience with natural language processing would also be desirable. SOFTWARE DEVELOPMENT ENVIRONMENTS: Should be familiar with MatLab, and program in C++ on PC's and Unix. Experience with Windows CE would also be desirable. Compensation will be competitive. INSTITUTION: CHCC at OGI is one of the world's leading institutions engaged in research on multimodal interaction. We are well-supported by DARPA, NSF, ONR, and other organizations. The QuickSet multimodal system is highly regarded and is a key part of numerous Government programs. To learn more about our Center and our work, please visit http://www.cse.ogi.edu/CHCC. AREA: Portland, Oregon is a fine place to live. There are numerous outdoors activities (skiing, windsurfing, hiking, beach combing, etc.) within 1 hour's drive. Yes, it rains here in winter (but it probably does something nasty where you live too.) However, one can often have 5 months of no/little rain in Spring/Summer/Fall. To apply, please send a resume, research statment, and names/contact information for at least 3 references to: Prof. Phil Cohen Center for Human-Computer Communication Oregon Graduate Institute of Science and Technology 20000 NW Walker Rd. Beaverton, OR 97006 503-690-1326 Fax: 503-690-1548 From Ronan.Reilly at ucd.ie Wed Oct 13 18:43:43 1999 From: Ronan.Reilly at ucd.ie (Ronan Reilly) Date: Wed, 13 Oct 1999 23:43:43 +0100 Subject: UCD, Ireland - Two-year graduate studentship Message-ID: <02aa01bf15cc$6a0de370$5f042b89@ucd.ie> Two-year graduate studentship on a computational neuroscience project Department of Computer Science, University College Dublin, Ireland The Cortical Software Re-Use (CSRU) project is a two-year interdisciplinary project involving the Departments of Mathematical Physics, Psychology, and Computer Science. Its aim is to apply a dynamical systems approach to modelling cortical computation with a specific focus on frontal lobe function. The Department of Computer Science at University College Dublin has funding of ?12,700 per year for two years for an MSc or PhD student to work on the project. The ideal candidate will have a background in computer science and some experience of computational modelling in the neuroscience area. For further details, contact Ronan.Reilly at ucd.ie. ________________________________ Ronan G. Reilly, PhD Department of Computer Science University College Dublin Belfield Dublin 4 Ireland voice: +353-1-706 2475 fax: +353-1-269 7262 e-mail: Ronan.Reilly at ucd.ie web: http://www.cs.ucd.ie/staff/rreilly/ From gbugmann at soc.plym.ac.uk Wed Oct 13 23:21:00 1999 From: gbugmann at soc.plym.ac.uk (Guido.Bugmann xtn 2566) Date: Thu, 14 Oct 1999 04:21:00 +0100 (BST) Subject: Postdoctoral Research Associate and one PhD student Message-ID: Dear Moderator, the annoucement below may be of interest to members of the list, the project involves learning (via instruction) and knowledge representation (probably connectionnist). Thanks in advance for posting this announcement. Guido Bugmann ----------------------------- Dr. Guido Bugmann Senior Research Fellow Centre for Neural and Adaptive Systems School of Computing University of Plymouth Plymouth PL4 8AA United Kingdom ----------------------------- Tel: (+44) 1752 23 25 66 / 41 Fax: (+44) 1752 23 25 40 Email: gbugmann at soc.plym.ac.uk or gbugmann at plymouth.ac.uk http://www.tech.plym.ac.uk/soc/Staff/GuidBugm/Bugmann.htm ----------------------------- Sorry if you receive multiple copies of this announcements. ----------------------------------------------------------- A Postdoctoral Research Associate and one PhD student are sought for an exciting project aimed at developing principles of "Instruction-Based Learning", i.e. teaching robots using natural language. This project explores a still-untapped method of knowledge acquisition and learning by intelligent systems: the acquisition of knowledge from Natural Language (NL) instruction. This is very effective in human learning and will be essential for adapting future intelligent systems to the needs of naive users. The aim of the project is to investigate real world Instruction Based-Learning (IBL) in a generic route instruction task. Users will engage in a dialogue with a mobile robot equipped with artificial vision, in order to teach it how to navigate a simplified maze-like environment. This experimental set-up will limit perceptual and control problems and also reduce the complexity of NL processing. The research will focus on the problem of how NL instructions can be used by an intelligent embodied agent to build a hierarchy of complex functions based on a limited set of low-level perceptual, motor and cognitive functions. We will investigate how the internal representations required for robot sensing and navigation can support a usable speech-based interface. Given the use of artificial vision and voice input, such a system can contribute to assisting visually impaired people and wheelchair users. This 3 year project for 2 research associates and one PhD student is funded by the UK Engineering and Physical Sciences Research Council. It is jointly managed by Dr. G. Bugmann of the Centre for Neural and Adaptive Systems of the University of Plymouth and by Dr. E. Klein of the Institute for Communicating and Collaborative Systems at the University of Edinburgh. The tasks are subdivided into two broad categories: (i) natural language processing and dialogue management at Edinburgh (1 RA) and (ii) robot experiments, knowledge representation and systems integration in Plymouth (1 RA, 1 RS). The project will involve sustained interaction and coordination between the two groups in order to support the required level of integration. The positions offered in Plymouth are: 1. Research Associate This position requires a strong background in the areas of systems architecture, knowledge representation, systems integration, robotics and artificial vision. Salary scale: 15205-19999 UK Pounds (Pay award pending.) 2. Research student The PhD student will examine some of the more speculative and open-ended aspects of IBL as a form of robot learning. Bursary: 6805 UK Pounds (funding for non-UK students submitted to restrictions). The two positions require skills in programming C++ and Perl/Python under Windows NT or LINUX. Experience with hardware issues and real world image processing is desirable. Applications should be received before 1 December, but the positions will remain open until suitable candidates are found. Candidates should send a CV, a description of motivations and the names of three references (by surface mail of email) to: Carole Watson School of Computing University of Plymouth Drake Circus Plymouth PL4 8AA, UK c.watson at plymouth.ac.uk Tel: +44 (0) 1752 23 25 41 Fax: +44 (0) 1752 23 25 40 Informal questions can be asked to: Dr. Guido Bugmann (abroad from 8 - 18 Oct.) gbugmann at soc.plym.ac.uk Tel: +44 (0) 1752 23 25 66 Fax: +44 (0) 1752 23 25 40 http://techweb.plym.ac.uk/soc/staff/guidbugm/bugmann.htm ----------------------------- From svensen at cns.mpg.de Thu Oct 14 03:11:09 1999 From: svensen at cns.mpg.de (Markus Svensen) Date: Thu, 14 Oct 1999 09:11:09 +0200 Subject: Post-doc, Medical Image Processing Message-ID: <3805820D.677F190A@cns.mpg.de> Dear Connectionists, please do not respond to me, but to Dr Frithjof Kruggel (details below). Markus Svensen RESEARCH POSITION Max-Planck-Institute of Cognitive Neuroscience Workgroup on Signal and Image Processing The Workgroup on Signal and Image Processing is seeking applicants for a Post-Doctoral Research Position in the area of image processing as applied to the analysis of anatomical data as revealed by magnetic resonance imaging (MRI). The successful applicant will be expected to participate in an advanced research and development effort centered on the following topics: o image segmentation of pathological MRI data o statistical analysis of characteristics of brain lesions o detection of changes in brain lesions with time Successful applicants should have an advanced degree in Mathematics, Physics, Computer Science or Medical Informatics, and significant experience in image processing as applied to medical problems. The term of the position is currently limited to one year, with a possible extension of another year. A starting date around January 2000 is expected. A competitive salary will be offered, with actual salary dependent on the applicant's qualifications and experience. We encourage qualified women to apply. Interested applicants should send a resume and a list of references to Dr. F. Kruggel (kruggel at cns.mpg.de) before October 31, 1999. Electronic application is preferred, but if necessary, applications may be sent to the address below. Dr. F. Kruggel Max-Planck-Institute of Cognitive Neuroscience Stephanstr. 1, 04103 Leipzig, Germany kruggel at cns.mpg.de Phone: (49) 341 9940 223, FAX: (49) 341 9940 221 From aweigend at stern.nyu.edu Thu Oct 14 07:10:03 1999 From: aweigend at stern.nyu.edu (Andreas Weigend) Date: Thu, 14 Oct 1999 07:10:03 -0400 (EDT) Subject: Jobs: Music Cognition / Machine Learning / Web Message-ID: <199910141110.HAA05655@sabai.stern.nyu.edu> A non-text attachment was scrubbed... Name: not available Type: text Size: 1359 bytes Desc: not available Url : https://mailman.srv.cs.cmu.edu/mailman/private/connectionists/attachments/00000000/16047f82/attachment.ksh From jagota at cse.ucsc.edu Thu Oct 14 14:04:16 1999 From: jagota at cse.ucsc.edu (Arun Jagota) Date: Thu, 14 Oct 1999 11:04:16 -0700 Subject: new e-survey in NCS Message-ID: <199910141804.LAA10346@sundance.cse.ucsc.edu> New refereed e-publication action editor: Yoshua Bengio W. Duch & N. Jankowski, Survey of Neural Transfer Functions Neural Computing Surveys 2, 163--212, 1999. 91 references. http://www.icsi.berkeley.edu/~jagota/NCS Abstract: The choice of transfer functions may strongly influence complexity and performance of neural networks. Although sigmoidal transfer functions are the most common, there is no /a priori/ reason why models based on such functions should always provide optimal decision borders. A large number of alternative transfer functions have been described in the literature. A taxonomy of activation and output functions is proposed, and advantages of various non-local and local neural transfer functions are discussed. Several lesser known types of transfer functions, and new combinations of activation/output functions are described. Universal transfer functions, parametrized to change from localized to delocalized type, are of greatest interest. Other types of neural transfer functions discussed here include functions with activations based on non-Euclidean distance measures, bicentral functions, formed from products or linear combinations of pairs of sigmoids, and extensions of such functions making rotations of localized decision borders in highly dimensional spaces practical. Nonlinear input preprocessing techniques are briefly described, offering an alternative way to change the shapes of decision borders. From stefan.wermter at sunderland.ac.uk Thu Oct 14 14:23:21 1999 From: stefan.wermter at sunderland.ac.uk (Stefan Wermter) Date: Thu, 14 Oct 1999 19:23:21 +0100 Subject: Neural networks based on neuroscience Message-ID: <38061F99.5444E062@sunderland.ac.uk> In September 1999 we held an EPSRC funded EmerNet workshop, an International Workshop on Emergent Neural Computational Architectures Based on Neuroscience. Topics in the Call for Neural Networks based on Neuroscience had included: 1.Synchronisation: How does the brain synchronise its processing? How does the brain schedule its processing? 2.Processing speed: How does the brain compute with relatively slow computing elements but still achieve rapid and real time performance? 3.Robustness: How does human memory manage to continue to operate despite failure of its components? 4.Modular construction: What can we learn from the brain for building modular more powerful artificial neural network architectures to solve larger tasks? 5.Learning in context: How can we build learning algorithms which consider context? How can we design incremental learning algorithms and dynamic architectures? This message is to announce that a summary as well as all the papers of this workshohp have been made available at: http://www.his.sunderland.ac.uk/emernet/icann99w.html Also at http://www.his.sunderland.ac.uk/emernet/ you may find additional related information on this topic best wishes, Stefan Wermter *************************************** Professor Stefan Wermter Research Chair in Intelligent Systems University of Sunderland Centre of Informatics, SCET St Peters Way Sunderland SR6 0DD United Kingdom phone: +44 191 515 3279 fax: +44 191 515 2781 email: stefan.wermter at sunderland.ac.uk http://www.his.sunderland.ac.uk/~cs0stw/ http://www.his.sunderland.ac.uk/ **************************************** From mlf at dlsi.ua.es Thu Oct 14 02:56:57 1999 From: mlf at dlsi.ua.es (Mikel L. Forcada) Date: Thu, 14 Oct 1999 08:56:57 +0200 Subject: Paper available Message-ID: <38057EB9.18293B09@dlsi.ua.es> Dear connectionists: The following paper, to appear in Neural Computation is now available via our web server: PDF: http://www.dlsi.ua.es/~mlf/stanc3web.pdf gzipped PostScript: http://www.dlsi.ua.es/~mlf/stanc3web.ps.gz Stable encoding of finite-state machines in discrete-time recurrent neural nets with sigmoid units Rafael C. Carrasco, Mikel L. Forcada, M. ?ngeles Vald?s-Mu?oz Departament de Llenguatges i Sistemes Inform?tics Universitat d'Alacant E-03071 Alacant (Spain) Ram?n P. ?eco Departament de Ci?ncies Experimentals i Tecnologia Universitat "Miguel Hern?ndez" E-03202 Elx (Spain) ABSTRACT In recent years, there has been a lot of interest in the use of discrete-time recurrent neural nets (DTRNN) to learn finite-state tasks, with interesting results regarding the induction of simple finite-state machines from input--output strings. Parallel work has studied the computational power of DTRNN in connection with finite-state computation. This paper describes a simple strategy to devise stable encodings of finite-state machines in computationally capable discrete-time recurrent neural architectures with sigmoid units, and gives a detailed presentation on how this strategy may be applied to encode a general class of finite-state machines in a variety of commonly-used first- and second-order recurrent neural networks. Unlike previous work that either imposed some restrictions to state values, or used a detailed analysis based on fixed-point attractors, the present approach applies to any positive, bounded, strictly growing, continuous activation function, and uses simple bounding criteria based on a study of the conditions under which a proposed encoding scheme guarantees that the DTRNN is actually behaving as a finite-state machine. -- _____________________________________________________________________ Mikel L. Forcada E-mail: mlf at dlsi.ua.es Departament de Llenguatges Phone: +34-96-590-3400 ext. 3384; i Sistemes Informatics also +34-96-590-3772. UNIVERSITAT D'ALACANT Fax: +34-96-590-9326, -3464 E-03071 ALACANT, Spain. ______________________________________________________________________ URL: http://www.dlsi.ua.es/~mlf ---------------------------------------------------------------------- From jpittman at microsoft.com Thu Oct 14 16:16:26 1999 From: jpittman at microsoft.com (Jay Pittman) Date: Thu, 14 Oct 1999 13:16:26 -0700 Subject: job posting at Microsoft Message-ID: <783D93998201D311B0CF00805FEAA07B010DF872@RED-MSG-42> Here's a job posting that may be of interest to the readers of this list. The Microsoft Handwriting Recognition team has immediate openings in handwriting recognition research and development. Responsibilities center on incorporating research into the Microsoft Handwriting Recognition (HR) engine. Experience in handwriting recognition technology is not required, but a solid background in Computer Science, Mathematics, or Statistics, is needed, as is the desire and ability to learn new techniques from textbooks or research papers. The HR engine is being developed for English and Japanese and includes components that implement neural nets, hidden Markov models, signal processing, statistical search, n-gram statistical language modeling, and context-free grammar language modeling. Improvements to the HR engine are targeted at increasing performance and accuracy, and improving usability problems for mainstream use in applications, such as note taking into future versions of Microsoft Office. Qualifications required include at least a Bachelors degree in Computer Science, Mathematics, EE/Signal Processing, or related field, and strong C/C++ coding skills. You must be able to code up your ideas to improve the recognizer in C/C++. A strong interest and ability to write and support your own code in C/C++ is crucial to being successful in the group. Send your resume, complete with references (if that is practical), to: Dr. James A. Pittman jpittman at microsoft.com I prefer email in either Word or ascii format, but you can also snailmail to: Dr. James A. Pittman, 30/3092 Microsoft Corporation One Microsoft Way Redmond, WA 98052 From ingber at ingber.com Thu Oct 14 19:34:44 1999 From: ingber at ingber.com (Lester Ingber) Date: Thu, 14 Oct 1999 18:34:44 -0500 Subject: More EEG data available Message-ID: <19991014183444.A3457@ingber.com> Since my previous announcement below, I was contacted by Padhraic Smyth who invited my to upload this data to the KDD Archive [http://kdd.ics.uci.edu] hosted by the Department of Information and Computer Science, University of California, Irvine. Working with the Librarian of this archive, Stephen Bay, I uploaded the full data set. Access is available via http://www.ingber.com/smni_eeg_data.html or http://kdd.ics.uci.edu/databases/eeg/eeg.html Lester : I have had several requests for EEG data since I had to remove access : from a large dataset I used for some of my studies, due to a disk crash : at an archive at U Oregon. : : I am making a small dataset (4 MB) and two larger datasets (40 MB : each) available at my Caltech archive. I do not know for how long I : will be able to keep the larger datasets available. I cannot keep the : largest dataset (700 MB) available there and do not plan on uploading : it anyplace else. All files sizes are in .gz-compressed formats. See : http://www.ingber.com/smni_eeg_data.html : : I hope you find this data useful for your own studies. -- Lester Ingber http://www.ingber.com/ PO Box 06440 Wacker Dr PO Sears Tower Chicago IL 60606-0440 http://www.alumni.caltech.edu/~ingber/ From steve at cns.bu.edu Sat Oct 16 08:03:54 1999 From: steve at cns.bu.edu (Stephen Grossberg) Date: Sat, 16 Oct 1999 08:03:54 -0400 Subject: A Neural Model of Motion Processing and Visual Navigation by Cortical Area MST Message-ID: The following article can be read at http://cns-web.bu.edu/Profiles/Grossberg/ A Neural Model of Motion Processing and Visual Navigation by Cortical Area MST, by Stephen Grossberg, Ennio Mingolla, and Christopher Pack. Cerebral Cortex, in press Cells in the dorsal medial superior temporal cortex (MSTd) process optic flow generated by self-motion during visually-guided navigation. A neural model shows how interactions between well-known neural mechanisms (log polar cortical magnification, Gaussian motion-sensitive receptive fields, spatial pooling of motion-sensitive signals, and subtractive extraretinal eye movement signals) lead to emergent properties that quantitatively simulate neurophysiological data about MSTd cell properties and psychophysical data about human navigation. Model cells match MSTd neuron responses to optic flow stimuli placed in different parts of the visual field, including position invariance, tuning curves, preferred spiral directions, direction reversals, average response curves, and preferred locations for stimulus motion centers. The model shows how the preferred motion direction of the most active MSTd cells can explain human judgments of self-motion direction (heading), without using complex heading templates. The model explains when extraretinal eye movement signals are needed for accurate heading perception, and when retinal input is sufficient, and how heading judgments depend on scene layouts and rotation rates. From harnad at coglit.ecs.soton.ac.uk Sat Oct 16 13:29:37 1999 From: harnad at coglit.ecs.soton.ac.uk (Stevan Harnad) Date: Sat, 16 Oct 1999 18:29:37 +0100 (BST) Subject: LANGUAGE EVOLUTION: 2 Psycoloquy Calls for Commentators Message-ID: (1) THE COGNITIVE PREREQUISITES FOR LANGUAGE (Burling) http://www.cogsci.soton.ac.uk/psyc-bin/newpsy?10.032 (2) LANGUAGE EVOLUTION AND THE COMPLEXITY CRITERION (Bichakjian) http://www.cogsci.soton.ac.uk/psyc-bin/newpsy?10.033 The two target articles whose abstracts follow below were published today in PSYCOLOQUY, a refereed journal of Open Peer Commentary sponsored by the American Psychological Association. Qualified professional biobehavioral, neural or cognitive scientists are hereby invited to submit Open Peer Commentary on either or both articles. Please email or consult the journal's websites below for Instructions if you are not familiar with format or acceptance criteria for PSYCOLOQUY commentaries (all submissions are refereed). To submit articles and commentaries or to seek information: EMAIL: psyc at pucc.princeton.edu URL: http://www.princeton.edu/~harnad/psyc.html http://www.cogsci.soton.ac.uk/psyc ----------------------------------------------------------------------- (1) THE COGNITIVE PREREQUISITES FOR LANGUAGE Target Article on Language-Prerequisites Robbins Burling Department of Anthropology 1020 LSA Building University of Michigan Ann Arbor MI 48109 USA rburling at umich.edu ABSTRACT: The first use of words by our early ancestors probably depended on four cognitive capacities: A rich conceptual understanding of the world around us; the ability to use and understand motivated signs, both icons and indices; the ability to imitate; the ability to infer the referential intentions of others. The latter three capacities are rare or absent in nonprimate mammals, but incipient in apes and well developed in modern humans. Before early humans could have begun to use words these capacities would have needed further development than is found in modern apes. It is not clear why selection favoured these skills more strongly in our ancestors than in the ancestors of apes. KEYWORDS: cognition; evolution; iconicity; imitation; language; names; theory-of-mind; words. Access full text at: http://www.cogsci.soton.ac.uk/psyc-bin/newpsy?10.032 ----------------------------------------------------------------------- (2) LANGUAGE EVOLUTION AND THE COMPLEXITY CRITERION Target Article on Language-Complexity Bernard H. Bichakjian Department of French University of Nijmegen, The Netherlands Bichakjian at let.kun.nl http://welcome.to/bichakjian ABSTRACT: Though it is increasingly accepted in the behavioral sciences, the evolutionary approach is still meeting resistance in linguistics. Linguists generally cling to the idea that alternative linguistic features are simply gratuitous variants of one another, while the advocates of innate grammars, who make room for evolution as a biological process, exclude the evolution of languages. The rationale given is that today's languages are all complex systems. This argument is based on the failure to distinguish between complexities of form and function. The proper analysis reveals instead that linguistic features have consistently decreased their material complexity, while increasing their functionality. A systematic historical survey will show instead that languages have evolved and linguistic features have developed along a Darwinian line. KEYWORDS: complexity, Indo-European, language evolution, lateralization, neoteny, word order. Access full text at: http://www.cogsci.soton.ac.uk/psyc-bin/newpsy?10.033 ----------------------------------------------------------------------- To submit articles and commentaries or to seek information: EMAIL: psyc at pucc.princeton.edu URL: http://www.princeton.edu/~harnad/psyc.html http://www.cogsci.soton.ac.uk/psyc From bap at cs.unm.edu Sun Oct 17 03:47:24 1999 From: bap at cs.unm.edu (Barak Pearlmutter) Date: Sun, 17 Oct 1999 01:47:24 -0600 (MDT) Subject: Positions Available Message-ID: POSTDOCTORAL, PREDOCTORAL and STAFF POSITIONS AVAILABLE in Blind Source Separation and Magnetoencephalography Prof Barak A. Pearlmutter Prof Akaysha C. Tang Dept of Computer Science Dept of Psychology Dept of Neurosciences Dept of Neurosciences University of New Mexico University of New Mexico http://www.cs.unm.edu/~bap/ http://kongzi.unm.edu Our growing Brain and Computation group is focusing most of its effort on functional brain imaging, magnetoencephalography in particular. We are applying blind source separation techniques to MEG data, and designing and conducting experiments that take advantage of MEG's unique combination of high temporal resolution and spatial sensitivity. The project spans the spectrum from experiment design and data acquisition to analysis and theory, and we have positions available at all levels for neuroscientists, computer scientists, physicists, differential geometers, etc who can contribute to - relevant theoretical work - developing source separation algorithms - implementing algorithms for processing brain imaging data - developing visualization tools and techniques - contributing to our open-source MEG software suite - designing and conducting experiments - collecting and analyzing data These positions are to be largely supported through funding from the new National Foundation for Functional Brain Imaging, http://www.nffbi.org/, which is supporting three centers (Minnesota, Mass Gen, and Albuquerque) and whose headquarters are located on the University of New Mexico campus. The lab enjoys a close relationship with the Albuquerque High Performance Computing Center, http://www.ahpcc.unm.edu/, through which we have access to superlative computational resources, and with the nearby Santa Fe Institute, http://www.santafe.edu/. We also have ties to the nearby VA Hospital, which hosts a 122-channel full-head neuromagnetometer. Applications on the Computer Science/Algorithms end of the spectrum should be sent to Barak A. Pearlmutter Computer Science Dept, FEC 313 University of New Mexico Albuquerque, NM 87131 bap at cs.unm.edu and those on the Neuroscience/Experimentation end of the spectrum to Akaysha C. Tang Department of Psychology, Logan Hall University of New Mexico Albuquerque, NM 87131 akaysha at kongzi.unm.edu Applications should include a full cv, names and contact information of three references, and a statement of research interests. The preferred format is emailed plain ascii, but emailed postscript or physical mail are also acceptable. Top candidates will be interviewed on-site. The University of New Mexico main campus is on historic Route 66 in Albuquerque, is an official botanical garden, and is seven minutes by car from a major airport and thirty from a ski resort. The state of New Mexico has the highest per-capita number of PhD's in the country, and offers outstanding outdoor recreational opportunities. For local information see links below http://kongzi.unm.edu/~akaysha/ or http://albuquerque.areaguides.net/. From harnad at coglit.ecs.soton.ac.uk Sun Oct 17 07:56:23 1999 From: harnad at coglit.ecs.soton.ac.uk (Stevan Harnad) Date: Sun, 17 Oct 1999 12:56:23 +0100 (BST) Subject: SELF-CONSCIOUSNESS/Bermudez: PSYC Call for Book Reviewers (844 l.) Message-ID: PSYCOLOQUY CALL FOR BOOK REVIEWERS Below is the Precis of "The Paradox of Self-Consciousness" by Jose Luis Bermudez (799 lines). This book has been selected for multiple review in PSYCOLOQUY. If you wish to submit a formal book review please write to psyc at pucc.princeton.edu indicating what expertise you would bring to bear on reviewing the book if you were selected to review it. (If you have never reviewed for PSYCOLOQUY or Behavioral & Brain Sciences before, it would be helpful if you could also append a copy of your CV to your inquiry.) If you are selected as one of the reviewers and do not have a copy of the book, you will be sent a copy of the book directly by the publisher (please let us know if you have a copy already). Reviews may also be submitted without invitation, but all reviews will be refereed. The author will reply to all accepted reviews. Full Psycoloquy book review instructions at: http://www.princeton.edu/~harnad/psyc.html http://www.cogsci.soton.ac.uk/psycoloquy/ Relevant excerpts: Psycoloquy reviews are of the book not the Precis. Length should be about 200 lines [c. 1800 words], with a short abstract (about 50 words), an indexable title, and reviewer's full name and institutional address, email and Home Page URL. All cited references that are electronically accessible should also have URLs indicated. AUTHOR'S RATIONALE FOR SOLICITING MULTIPLE BOOK REVIEW: The book offers a novel approach to the study of self-consciousness, integrating philosophical argument with detailed study of empirical work from a range of disciplines. It provides a framework for linking together distinct areas of cognitive science which are rarely discussed together and discusses some fundamental problems in the foundations of psychology (such as the nature of concepts and the possibility of thought without language). I am continuing to work on some of the central themes of the book and would greatly benefit from feedback from the biobehavioral and cognitive science community. ----------------------------------------------------------------------- psycoloquy.99.10.035.self-consciousness.1.bermudez Sun Oct 17 1999 ISSN 1055-0143 (47 paragraphs, 30 references, 799 lines) PSYCOLOQUY is sponsored by the American Psychological Association (APA) Copyright 1999 Jose Luis Bermudez THE PARADOX OF SELF-CONSCIOUSNESS (REPRESENTATION AND MIND) Precis of Bermudez on Self-Consciousness [MIT/Bradford, 1998 xiv, 236 pp. ISBN: 0-262-02441-1] Jose Luis Bermudez Department of Philosophy University of Stirling Stirling FK9 4LA Scotland CREA Ecole Polytechnique 1 Rue Descartes 75005 Paris France jose.bermudez at stir.ac.uk ABSTRACT: This book addresses two fundamental questions in the philosophy and psychology of self-consciousness: (1) Can we provide a noncircular account of full-fledged self-conscious thought and language in terms of more fundamental capacities? (2) Can we explain how full-fledged self-conscious thought and language can arise in the normal course of human development? I argue that a paradox (the paradox of self-consciousness) arises from the apparent strict interdependence between self-conscious thought and linguistic self-reference. Responding to the paradox, I draw on recent work in empirical psychology and philosophy to cut the tie between self-conscious thought and linguistic self-reference. The book studies primitive forms of nonconceptual self-consciousness manifested in visual perception, somatic proprioception, spatial reasoning and interpersonal psychological interactions. KEYWORDS: cognitive maps; concepts; content; ecological self; navigation; proprioception; self-consciousness; self-reference; visual perception; I. INTRODUCTION 1. Philosophy and the neurosciences have an uneasy relationship. Fruitful engagement is rare in either direction. This is partly the inevitable result of the division of academic labours. But there is also a deeper reason. The dominant methodological conception governing work in the cognitive sciences involves a distinction of levels of explanation. Marr's theory of vision has often been held up as a model which the cognitive sciences in general ought to follow mainly, of course, because it is one of the very few worked out and satisfying theoretical treatments of a cognitive capacity that cognitive science has so far produced. As is well-known, Marr's approach to the study of the visual system is top-down (Marr 1982). He starts with an abstract specification of the functional tasks that the visual system has to perform, hypothesises a series of algorithms that could compute these functional tasks and then speculates about the implementation of those algorithms at the neural level. Each of the levels of explanation at which the theory operates is relatively autonomous, although of course the computational level models the realisation of the functions identified at the functional level and the implementational level explains how the functions identified at the computational level are realized. The resulting theory is, of course, a dazzling achievement. But there are hidden implications in taking it as a general paradigm for cognitive science. Taking it as a paradigm makes it natural to think, for example, that the place of philosophy is at the functional level and, correspondingly, that the place of cognitive neuroscience is at the implementational level. The result, of course, is that the two disciplines are effectively insulated from each other by the intervening computational level of explanation. 2. There is an obvious problem, however, with generalizing Marr's approach. The problem is that the distinction of levels of explanation really makes sense only where one can identify a clear functional task or set of tasks that need to be carried out. But it is not clear that this can be done outside the restricted domain of encapsulated modules such as the early visual system, the language-parsing system or the face recognition system. Fodor, the most articulate defender of this methodological approach, has clearly appreciated this, and drawn the drastic conclusion that cognitive science cannot hope to shed any light on the so-called central processes of cognition. A more sensible lesson to draw, I think, is that outside this restricted domain a more interactive conception of the relation between the levels of explanation is appropriate. There must be constraints on theorizing at the functional and computational levels. On the top-down approach these constraints emerge from clearly defined functional tasks. But where there are no such functional tasks explanation cannot be purely top-down. There must be constraints and programmatic suggestions moving in both directions. 3. The difficulty in putting this programme into practice is identifying the points of contact between neuroscientific concerns and, for example, philosophical concerns. In this prcis of my book The Paradox of Self-Consciousness (Bermdez 1998) I identify some of the key areas where neuroscientific and philosophical issues intersect in the study of self-consciousness, a form of cognition about as far as it is possible to get from the encapsulated modules where top-down analyses can be so profitably applied. II. THE PARADOX OF SELF-CONSCIOUSNESS 4. In thinking about self-consciousness we need to start with the phenomenon of first-person thought. Most, if not all, of the higher forms of self-consciousness presuppose our capacity to think about ourselves. Consider, for example, self-knowledge, the capacity for moral self- evaluation and ability to construct a narrative of our past. Although much of what we think when we think about ourselves involves concepts and descriptions also available to us in our thoughts about other people and other objects, our thoughts about ourselves also involve an ability that we cannot put to work in thinking about other people and things - namely, the ability to apply those concepts and descriptions uniquely to ourselves. I shall follow convention in referring to this as the capacity to entertain 'I'-thoughts. 5. 'I'-thoughts of course involve self-reference, but it is self-reference of a distinctive kind. Consider the following two ways in which I might entertain thoughts that refer to myself: (1) JLB thinks: JLB is about to be attacked by a poisonous spider (2)JLB thinks: I am about to be attacked by a poisonous spider It is clear that these are very different thoughts, even though they are both thoughts about the same person, namely me. Even if I am suffering a temporary attack of amnesia that has led me to forget my own name I can think the first thought with equanimity. Not so the second. 6. This property of I-thoughts is sometimes described as their immunity to error through misidentification, where this means (roughly) that one cannot think an 'I'-thought without knowing that it is in fact about oneself (Shoemaker 1968, Evans 1982). This feature of 'I'-thoughts is closely tied to the well-known linguistic property of the first-person pronoun, namely, that the first-person pronoun I always refers to the person uttering it. 7. Putting these two properties together suggests the following deflationary account of self- consciousness: (A) Once we have an account of what it is to be capable of thinking 'I'-thoughts we will have explained everything that is distinctive about self-consciousness. (B) Once we have an account of what it is to be capable of thinking thoughts that are immune to error through misidentification we will have explained everything that is distinctive about the capacity to think 'I'-thoughts. (C) Once we have explained what it is to master the semantics of the first person pronoun (e.g. via mastery of some version of the token-reflexive rule that a given utterance of I always refers to the person uttering it), we will have explained everything that is distinctive about the capacity to think thoughts that are immune to error through misidentification. 8. The problem with the deflationary view is that first-person self-reference is itself dependent upon 'I'-thoughts in a way that creates two forms of vicious circularity which collectively I term the paradox of self-consciousness The first type of circularity (explanatory circularity), arises because the capacity for self-conscious thought must be presupposed in any satisfactory account of mastery of the first person pronoun. I cannot refer to myself as the producer of a given token of 'I' without, for example, knowing that I intend to refer to myself - which is itself a self-conscious thought of the type that we are trying to explain. The second type of circularity (capacity circularity) arises because this interdependence rules out the possibility of explaining how the capacity either for self-conscious thought or for linguistic mastery of the first person pronoun arises in the normal course of human development. It does not seem possible to meet the following constraint: The Acquisition Constraint If a given psychological capacity is psychologically real then there must be an explanation of how it is possible for an individual in the normal course of human development to acquire that capacity. Neither self-conscious thought nor linguistic mastery of the first person pronoun is innate, and yet each presupposes the other in a way that seems to imply that neither can be acquired unless the other capacity is already in place. III. ESCAPING THE PARADOX OF SELF-CONSCIOUSNESS 9. The strategy that I employ in the book to escape the paradox of self-consciousness involves making a clear distinction between (a) those forms of full-fledged self-consciousness which presuppose mastery of the first person concept and linguistic mastery of the first person pronoun, and (b) those forms of primitive or nonconceptual self-consciousness which do not require any such linguistic or conceptual mastery. It is these nonconceptual forms of self-consciousness that allow us to escape both the types of circularity Ive just identified. I identified such nonconceptual forms of self-awareness in four domains: (1) perceptual experience (2) somatic proprioception (bodily self-awareness) (3) self-world dualism in spatial reasoning (4) psychological interaction 10. The basic result is that the domain of self-consciousness is far wider than it has been held to be by philosophers. Self-consciousness has often been thought to be the highest form of human cognition, and many philosophers, famous and not so famous, have correspondingly thought that a philosophical account of self-consciousness would be the Archimedean point for a satisfactory account of human thought. But the premise is flawed. Self-consciousness is something we share with prelinguistic infants and with many members of the animal kingdom. The highly conceptual forms of self-consciousness emerge from a rich foundation of nonconceptual forms of self- awareness. As I will try to bring out, recognising this builds a bridge between philosophical interests and neuroscientific ones. IV. THE SELF OF ECOLOGICAL OPTICS 11. One of J. J. Gibsons great insights in the study of visual perception was that the very structure of visual perception contains propriospecific information about the self, as well as exterospecific information about the distal environment (Gibson 1979). Visual perception involves self-perception at the same time as it involves perception of the world. This is the most primitive form of nonconceptual self-awareness, the foundation on which all other forms of self-awareness are built. 12. Gibson stresses certain peculiarities of the phenomenology of the field of vision. Notable among these is the fact that the field of vision is bounded. Vision reveals only a portion of the world to the perceiver at any given time (roughly half in the human case, due to the frontal position of the eyes). The boundedness of the field of vision is part of what is seen, and the field of vision is bounded in a way quite unlike the way in which spaces are bounded within the field of vision. The self appears in perception as the boundary of the visual field a moveable boundary that is responsive to the will. 13. The boundedness of the visual field is not the only way in which the self becomes manifest in visual perception. The field of vision contains other objects that hide, or occlude, the environment. These objects are, of course, various parts of the body. The nose is a particularly obvious example, so distinctively present in just about every visual experience. The cheekbones, and perhaps the eyebrows, occupy a slightly less dominant position in the field of vision. And so too, to a still lesser extent, do the bodily extremities, hands, arms, feet and legs. They protrude into the field of vision from below in a way that occludes the environment, and yet which differs from the way in which one non-bodily physical object in the field of vision might occlude another. They are, as Gibson points out, quite peculiar objects. All objects, bodily and non-bodily, can present a range of solid angles in the field of vision (where by a solid angle is meant an angle with its apex at the eye and its base at some perceived object), and the size of those angles will of course vary according to the distance of the object from the point of observation. The further away the object is, the smaller the angle will be. This gives rise to a clear, and phenomenologically very salient, difference between bodily and non-bodily physical objects. The solid angles subtended by occluding body-parts cannot be reduced below a certain minimum. Perceived body-parts are, according to Gibson, 'subjective objects' in the content of visual perception. 14. But these self-specifying structural invariants provide only a fraction of the self-specifying information available in visual perception. 15. The mass of constantly changing visual information generated by the subjects motion poses an immense challenge to the perceptual systems. How can the visual experiences generated by motion be decoded so that subjects perceive that they are moving through the world? Gibsons notion of visual kinesthesis is his answer to this traditional problem. Whereas many theorists have assumed that motion perception can only be explained by the hypothesis of mechanisms which parse cues in the neutral sensations into information about movement and information about static objects, the crucial idea behind visual kinesthesis is that the patterns of flow in the optic array and the relations between the variant and invariant features make available information about the movement of the perceiver, as well as about the environment. 16. As an example of such a visually kinesthetic invariant, consider that the optical flow in any field of vision starts from a centre, that is itself stationary. This stationary centre specifies the point that is being approached, when the perceiver is moving. The aiming point of locomotion is at the vanishing point of optical flow. 17. Striking experiments have brought out the significance of visual kinesthesis. In the so-called moving-room experiments, subjects are placed on the solid floors of rooms whose walls and ceilings can be made to glide over a solid and immoveable floor (Lishman and Lee 1973). If experimental subjects are prevented from seeing their feet and the floor is hidden, then moving the walls backwards and forwards on the sagittal plane creates in the subjects the illusion that they are moving back and forth. This provides strong support for the thesis that the movement of the perceiver can be detected purely visually, since visual specification of movement seems to be all that is available. An even more striking illustration emerges when young children are placed in the moving room, because they actually sway and lose their balance (Lee and Aronson 1975). 18. The theory of ecological optics identifies a third form of self-specifying information existing in the field of vision. This is due to the direct perception of a class of higher-order invariants which Gibson terms affordances. It is in the theory of affordances that we find the most sustained development of the ecological view that the fundamentals of perceptual experience are dictated by the organism's need to navigate and act in its environment. The uncontroversial premise from which the theory of affordances starts is that objects and surfaces in the environment have properties relevant to the abilities of particular animals, in virtue of which they allow different animals to act and react in different ways. 19. According to Gibson, information specifying affordances is available in the structure of light to be picked up by the creature as it moves around the world. The possibilities which the environment affords are not learnt through experience, and nor are they inferred. They are directly perceived as higher-order invariants. And of course, the perception of affordances is a form of self-perception - or, at least, a way in which self-specifying information is perceived. The whole notion of an affordance is that of environmental information about ones own possibilities for action and reaction. 20. Recognising the existence of the ecological self, as it has come to be known (Neisser 1988), is the first step in resolving the paradox of self-consciousness. It removes the need to explain how infants can bootstrap themselves into the first-person perspective. The evidence is overwhelming that nonconceptual first person contents are available more or less from the beginning of life. Illustrations are to be found in: (1) neonatal distress crying (Martin and Clark 1982) (2) neonatal imitation (Meltzoff and Moore 1977) (3) infant reaching behaviour (Field 1976, Von Hofsten 1982) (4) visual kinesthesis (Lee and Aronson 1984, Butterworth and Hicks 1977, Pope 1984) Let me turn now to some more bottom-level concerns. 21. When we move to considering the neural underpinnings of this form of self-specifying information in visual perception we move into an area that has been fairly closely studied by neuroscientists and experimental psychologists. Particularly relevant here is the proposal, currently under much discussion, that there are two distinct cortical pathways in the human visual system, each carrying distinct types of information (Ungerleider and Mishkin 1982, Goodale and Milner 1992). The distinction between the information carried by the dorsal (infero-temporal) and the ventral (occipito-temporal) pathways respectively has been conceptualized in different ways. Mishkin and Ungerleider see it as a distinction between information about the spatial relations in which an object might stand to the perceiver and information that allows the recognition of objects. Goodale and Milner, in contrast, take the distinction to be between recognitional information about the intrinsic properties of objects (eg their colour, shape and so on) and visuo-motor information about the extrinsic properties of objects (eg their spatial position, orientation, height and so forth). 22. It has been suggested that the action-based self-specifying information that Gibson discusses at the phenomenological level in terms of affordances and invariants in optical flow seems to be carried in the ventral stream (McCarthy 1993). It is far from clear to me, however, that Gibsons insights into the blend of propriospecific and exterospecific information in visual perception fits at all neatly into the proposed distinction of pathways, whether as construed by Mishkin and Ungerleider or by Goodale and Milner at least, if we assume that those processing distinctions are supposed to mark a distinction at the level of conscious phenomenology. The basic concept of an affordance seems to straddle the distinction between where and what, or between recognition and pragmatic. Interestingly, this scepticism about the phenomenological significance of the two cortical pathways is supported by recent work which suggests that the two visual pathways actually collaborate in the control of action (Jeannerod 1997). V. SOMATIC PROPRIOCEPTION AND THE BODILY SELF 23. Gibson's insights into the structure of visual perception were partly vitiated by his insistence on downplaying the importance of somatically derived information about the self. Visual kinesthesis and the perceptual invariants stressed by Gibson are adequate for distinguishing self-movement from movement of the environment, but they are unable to distinguish passive self-movement from active self-movement. They can inform the subject of his movement relative to the environment, but (crudely speaking) they do not tell him whether or not he is moving under his own steam. A different form of self-awareness is required at this point the bodily self-awareness of proprioception. 24. One particularly vivid illustration of the importance of these forms of proprioceptive information comes from the documented cases of complete deafferentation patients who have effectively lost all bodily sensation, either from below the neck in the case of Jonathan Coles patient IW or from below the jaw in Jacques Paillards patient GL (Cole and Paillard 1995). Although IW, unlike GL, can walk, everything he does has to be performed under visual control. Without visual feedback he is incapable of orienting himself and acting. So much so that he sleeps with the light on - if he woke up in the dark he would have no idea where his body was and would never be able to find the light switch. It is interesting, furthermore, to watch a video of him walking. His head is bent forward and pointing downwards so that he can keep his legs and feet in sight constantly. 25. There is a popular sense of self-conscious on which IW seems to be more self-conscious than we are, for the simple reason that everything he does requires his full attention. But this is not the sense of self-consciousness in which I am interested. What is striking about deafferented subjects is how the subjective sense of the body as a bounded spatial entity responsive to the will collapses in the absence of somatic proprioception and can only be partially reestablished with great artificiality and great difficulty. IW and GL are self-conscious in the popular sense precisely because they fail to be self-conscious in a more primitive and fundamental sense. 26. What is this more primitive and fundamental form of self-consciousness that we derive from somatic proprioception? It seems to me to have a tripartite structure. In exploring this it will be useful to start with a list of the principal types of proprioceptive information and their physiological sources. The following is taken from the general introduction to Bermdez, Marcel and Eilan 1995: Information about pressure, temperature and friction from receptors on the skin and beneath its surface. Information about the state of joints from receptors in the joints, some sensitive to static position, some to dynamic information. Information about balance and posture from the vestibular system in the inner ear; the head/trunk dispositional system; and information from pressure on any parts of the body that might be in contact with a gravity-resisting surface. Information about bodily disposition and volume obtained from skin-stretch. Information about nutritional and other homeostatic states from receptors in the internal organs. Information about muscular fatigue from receptors in the muscles. Information about general fatigue from cerebral systems sensitive to blood composition. Information about bodily disturbances derived from nociceptors. 27. At the simplest level, somatic proprioception is a form of self-consciousness simply in virtue of providing information about the embodied self. This is not particularly interesting, although it is worth noting that proprioception gives information about the embodied self that is immune to error through misidentification in the sense discussed earlier. It cannot be the case that one receives proprioceptive information without being aware that the information concerns ones own body. 28. More importantly, somatic proprioceptive information provides a way, perhaps the most primitive way, of registering the boundary between self and non-self. To appreciate this we need to note that there is an important variation among these somatic information systems vary along several dimensions. Some provide information solely about the body (eg. the systems providing information about general fatigue and nutrition). The vestibular system, in contrast, is concerned with bodily balance and hence with the relation between the body and the environment. Other systems can be deployed to yield information either about the body or about the environment. Receptors in the hand sensitive to skin stretch, for example, can provide information about the hand's shape and disposition at a time, or about the shape of small objects. Similarly, receptors in joints and muscles can yield information about how the relevant limbs are distributed in space, or, through haptic exploration, about the contours and shape of large objects. 29. These latter information systems, underpinning the sense of touch, yield a direct sense of the limits of the body and hence of the limits of the self. This is one step further in the development of what might be termed self-world dualism than comes with the self-specifying information in visual perception. The self of visual perception, the ecological self, is schematic and geometrical. Its properties are purely spatial, defined by patterns in the optical flow. It is only in virtue of the sense of touch that the body is experienced as a solid and bounded entity in the world. 30. It is known that a somatotopic map of the surface of the body exists in the somatosensory cortex, and it is natural to think that this plays a key role in subserving the registration of the boundary between self and non-self. Some confirmation confirmation for this will be found in the fascinating work that has been done by V. S. Ramachandran (1994) on somatosensory remapping to explain the well-documented phenomenon of referred sensations in amputees experiencing phantom limbs. The felt boundaries of the body can change as the area in the Penfield homunculus that formerly received input from the amputated limb is invaded by sensory input from nearby areas. 31. The final feature of proprioceptive self-awareness extends this sense of the body as an object. Through feedback from kinesthesia, joint-position sense and the vestibular system we become aware of the body as an object responsive to the will. Proprioception gives us a sense, not just of the embodied self as spatially extended and bounded, but also as a potentiality for action. 32. In this context it might be helpful to point to the role of proprioceptively derived information in the construction of the cross-modal egocentric space within which action takes place. It is well- known that lesions to the posterior parietal cortex produce spatial deficits in primates, human and non-human, and the inference frequently drawn is that the posterior parietal cortex is the brain area where the representation of space is computed. Recent neurophysiological work based on recordings from single neurons has suggested that the distinctive contribution of the posterior parietal cortex is the integration of information from various modalities to generate coordinate systems. Information about visual stimuli is initially transmitted in retinal coordinates. Calibrating this with information about eye position yields head-centred coordinates and further calibration with proprioceptively-derived information yields a body-centred frame of reference. The distal targets of reaching movements are encoded on this modality-free frame of reference, as are motor commands. VI. POINTS OF VIEW 33. The nonconceptual first person contents implicated in somatic proprioception and the pick-up of self-specifying information in visual perception provide very primitive forms of nonconceptual self-consciousness, albeit ones that can plausibly be viewed as in place from birth or shortly afterwards. A solution to the paradox of self-consciousness, however, requires showing how we can get from these primitive forms of self-consciousness to the fully-fledged self-consciousness that comes with linguistic mastery of the first person pronoun. This progression will have to be both logical (in a way that will solve the problem of explanatory circularity) and ontogenetic (in a way that will solve the problem of capacity circularity). Clearly, this requires that there be forms of self- consciousness which, while still counting as nonconceptual, are nonetheless more developed than those yielded by somatic proprioception and the structure of exteroceptive perception and, moreover, that it be comprehensible how these more developed forms of nonconceptual self- consciousness should have 'emerged' out of basic nonconceptual self-consciousness. 34. The dimension along which forms of self-consciousness must be compared is the richness of the conception of the self which they provide. Nonetheless, a crucial element in any form of self- consciousness is the way in which it makes possible for the self-conscious subject to distinguish between self and environment what many developmental psychologists term self-world dualism. In this sense self-consciousness is essentially a contrastive notion. One implication of this is that a proper understanding of the richness of the conception of the self which a given form of self- consciousness provides requires taking into account the richness of the conception of the environment with which it is contrasted. In the case both of somatic proprioception and of the pick- up of self-specifying information in exteroceptive perception, there is a relatively impoverished conception of the self associated with a comparably impoverished conception of the environment. One prominent limitation is that both are synchronic rather than diachronic. The distinction between self and environment that they offer is a distinction that is effective at a time but not over time. The contrast between propriospecific and exterospecific invariants in visual perception, for example, provides a way in which a creature can distinguish between itself and the world at any given moment, but this is not the same as a conception of oneself as an enduring thing distinguishable over time from an environment which also endures over time. 35. To capture this diachronic form of self-world dualism I introduced the notion of a nonconceptual point of view. Having a nonconceptual point of view on the world involves taking a particular route through the environment in such a way that one's perception of the world is informed by an awareness that one is taking such a route. This diachronic awareness that one is taking a particular route through the environment turned out to involve two principal components a non-solipsistic component and a spatial awareness component. 36. The non-solipsistic component is a subject's capacity to draw a distinction between his experiences and what those experiences are experiences of, and hence his ability to grasp that an object exists at times other than those at which it is experienced. This requires the exercise of recognitional abilities involving conscious memory and can be most primitively manifested in the feature-based recognition of places. This is the beginning of an understanding of the world as an articulated, structured entity. 37. The spatial awareness component of a nonconceptual point of view can be glossed in terms of possession of an integrated representation of the environment over time an understanding not just of how the articulated components of the external world fit together spatial, but also of the perceivers own spatial location in the world as a moving perceiver and agent. 38. That a creature possesses such an integrated representation of the environment is manifested in three central cognitive/navigational capacities: The capacity to think about different routes to the same place The capacity to keep track of changes in spatial relations between objects caused by its own movements relative to those objects The capacity to think about places independently of the objects or features located at those places. Powerful evidence from both ethology and developmental psychology indicates that these central cognitive/navigational capacities are present in both nonlinguistic and prelinguistic creatures. 39. This conception of a nonconceptual point of view provides a counterbalance to some important recent work on animal representations of space and their neurophysiological coding. Chapters 5 and 6 of Gallistels The Organization of Learning defend the thesis that all animals from insects upwards deploy cognitive maps with the same formal characteristics in navigating around the environment. Gallistel argues that the cognitive maps that control movement in animals all preserve the same set of geometric relations within a system of earth-centred (geocentric) coordinates. These relations are metric relations. The distinctive feature of a metric geometry is that it preserves all the geometric relations between the points in the coordinate system. Gallistel's thesis is that, although the cognitive maps of lower animals have far fewer places on them, they record the same geometrical relations between those points as humans and other higher animals. Moreover, he offers a uniform acount of how such metric cognitive maps are constructed in the animal kingdom. Dead reckoning (the process of keeping track of changes in velocity over time) yields an earth-centred representation of vantage points and angles of view which combines with current perceptual experience of the environment to yield an earth-centred cognitive map. 40. Without, of course, wishing to challenge Gallistels central thesis that all animal cognitive maps from insects up preserve geometric relations, it nonetheless seems wrong to draw the conclusion that all animals represent space in the same way. Just as important as how animals represent spatial relations between objects is how they represent their own position within the object-space thus defined. And it is here, in what we should think of as not just their awareness of space but also their awareness of themselves as spatially located entities, that we see the major variations and the scale of gradations that the theorists whom Gallistel is criticising have previously located at the level of the cognitive map. VII. PSYCHOLOGICAL SELF-AWARENESS 41. Possession of a nonconceptual point of view manifests an awareness of the self as a spatial element moving within, acting upon and being acted upon by the spatial environment. This is far richer than anything available through either somatic proprioception or the self-specifying information available in exteroceptive perception. Nonetheless, like these very primitive forms of self-consciousness, a nonconceptual point of view is largely awareness of the material self as a bearer of physical properties. This limitation raises the question of whether there can be a similarly nonconceptual awareness of the material self as a bearer of psychological properties. 42. There appear to be three central psychological properties defining the core of the concept of a psychological subject the property of being a perceiver, the property of being an agent, and the property of being a bearer of reactive attitudes. Research on the social cognition of infants shows that there are compelling grounds for attributing to prelinguistic infants in the final quarter of the first year awareness of themselves as bearers of all three of these properties. 43. Psychological self-awareness as a perceiver is manifested in the phenomenon of joint selective visual attention, where infants (a) attend to objects as a function of where they perceive the attention of others to be directed (Scaife and Bruner 1975, Bruner 1975), and (b) direct another individuals gaze to an object in which they are interested (Leung and Reinhold 1981, Stern 1985). In (b), for example, the infant tries to make the mother recognise that he, as a perceiver, is looking at a particular object, with the eventual aim that her recognition that this is what he is trying to do will cause the mother to look in the same direction. 44. Psychological self-awareness as an agent is manifested in the collaborative activities that infants engage in with their care-givers (coordinated joint engagement). Longitudinal studies (e.g. Trevarthen and Hubley 1978) show infants not just taking pleasure in their own agency (in the way that many infants show pleasure in the simple ability to bring about changes in the world, like moving a mobile), but also taking pleasure in successfully carrying out an intention - a form of pleasure possible only for creatures aware of themselves as agents. When, as it frequently is, the intention successfully carried out is a joint intention, the pleasure shared with the other participants reflects an awareness that they too are agents. 45. Psychological self-awareness as a bearer of reactive attitudes is apparent in what developmental psychologists call social referencing (Klinnert et al. 1983). This occurs when infants regulate their own behaviour by investigating and being guided by the emotional reactions of others to a particular situation. The infants willingness to tailor his own emotional reactions to those of his mother presuppose an awareness that both he and she are bearers of reactive attitudes. VIII. CONCLUSION 46. The four types of primitive or nonconceptual self-awareness provide the materials for resolving the paradox of self-consciousness. On the one hand, the problem of capacity circularity can be blunted by showing how it is conceivable that the capacity for full-fledged, conceptual self- consciousness could emerge from the basis of the primitive forms of self-consciousness discussed. On the other, the problem of explanatory circularity can be solved by giving an account of what it is to have mastery of the first-person pronoun that shows how the first-person thoughts involved can be understood at the nonconceptual level. 47. Instead of going into the details of how either of these goals can be achieved, I would like to return to the methodological reflections with which I began. I sketched out what I take to be a dominant approach to the methodology of cognitive science the top-down approach that clearly distinguishes the functional, computational and implementational levels of explanation. As I suggested, this approach really seems applicable only where there are clearly defined identifiable, functional tasks, and consequently is only going to work for peripheral rather than central cognitive processes. The corollary, as Fodor has clearly seen, is that we can expect little illumination of central processes from the cognitive sciences. What Ive tried to sketch out is an alternative approach, one where the distinction of levels of explanation does not correspond to a division of explanatory labour. Ive explored how attending to a particular philosophical puzzle about self- consciousness, perhaps the paradigm central cognitive process, brings out the importance of forms of self-consciousness that look as if they can only be understood by a more interactive collaboration between disciplines whose spheres of competence are so clearly separated on the conventional view. REFERENCES: Bermudez, J. L. 1998. The Paradox of Self-Consciousness. Cambridge MA. MIT Press. Bermudez, J. L., Marcel, A. J. and Eilan, N. (Eds.) 1995. The Body and the Self. Cambridge MA. MIT Press. Bruner, J. S. 1975. 'The ontogenesis of speech acts' in Journal of Child Language 2, 1-19. Butterworth, G. E., and Hicks, L. 1977. 'Visual proprioception and postural stability in infancy: A Developmental Study' in Perception 6, 255-262. Cole, J. and Paillard, J. 1995. 'Living without touch and peripheral information about body position and movement: Studies with deafferented subjects' in Bermudez, Marcel and Eilan (Eds.) 1995. Eilan, N., McCarthy, R. and Brewer, M. W. (Eds.).1993. Spatial Representation: Problems in Philosophy and Psychology. Oxford. Basil Blackwell. Evans, G. 1982. The Varieties of Reference. Oxford. Clarendon Press. Field, J. 'Relation of young infants' reaching behaviour to stimulus distance and solidity' in Developmental Psychology 12, 444-448. Fodor, J. 1983. The Modularity of Mind. Cambridge MA. MIT Press. Gallistel, C. R. 1990. The Organization of Learning. Cambridge MA. MIT Press. Gibson, J. J. 1979. The Ecological Approach to Visual Perception. Boston. Houghton Mifflin. Goodale, M. A. and Milner, A. D. 1992. 'Separate visual pathways for perception and action'. Trends in Neuroscience 15, 20-25. Jeannerod, M. 1997. The Cognitive Neuroscience of Action. Oxford. Basil Blackwell. Klinnert, M. D., Campos, J. J., Sorce, J. F. Emde, R. N. Svejda, M. 1983. 'Emotions as behaviour regulators: Social referencing in infancy' in Plutchik and Kellerman 1983. Lee, D. N., and Aronson, E. 1974. 'Visual proprioceptive control of standing in human infants' in Perception and Psychophysics 15: 529-532. Leung, E. and Rheinhold, H. 1981. 'Development of pointing as a social gesture' in Developmental Psychology 17, 215-220. Lishman, J. R., and Lee, D. N. 1973. 'The autonomy of visual kinaesthetics' in Perception 2: 287-94. Mccarthy, R. A. 1993. 'Assembling Routines and Addressing Representations: An Alternative Conceptualization of 'What' and 'Where' in the Human Brain' in Eilan, Mccarthy and Brewer (Eds.) 1993. Marr, D. 1982. Vision. San Fransisco. W. H. Freeman. Martin, G. B. and Clark, R. D. 1982. Distress Crying in Neonates: Species and Peer Specificity in Developmental Psychology 18, 3-9. Meltzoff, A. N. and Moore, M. K. 1977. 'Imitation of facial and manual gestures by human neomnates' in Science 198, 75-78. Neisser, U. 1988. 'Five Kinds of Self-Knowledge' in Philosophical Psychology, 1, 35-59. Pope, M. J. 1984. Visual Proprioception in Infant Postural Development. PhD Thesis. University of Southampton. Ramachandran, V. S. 1994. Phantom Limbs, Neglect Syndromes, Repressed Memories, and Freudian Psychology in International Review of Neurobiology 37, 291-333. Scaife, M. and Bruner, J. S. 1975. 'The capacity for joint visual attention in the infant' in Nature 253, 265-266. Shoemaker, S. 1968. 'Self-reference and self-awareness' in The Journal of Philosophy 65, 555-567. Stern, D. 1985. The Interpersonal World of the Infant. New York. Basic Books. Trevarthen, C. and Hubley, P. 1978. 'Secondary Intersubjectivity: Confidence, Confiding and Acts of Meaning in the First Year' in LOCK (Ed.) 1978. Ungerleider, M. and Mishkin, L. 1982. Two cortical visual systems. In D. J. Ingle, M. A. Goodale and R. J. W. Mansfield (Eds.), Analysis of Visual Behaviour. Cambridge MA. MIT Press. Von Hofsten, C. 1982. 'Foundations for Perceptual Development' in Advances in Infancy Research 2, 241-261. From harnad at coglit.ecs.soton.ac.uk Sun Oct 17 09:12:56 1999 From: harnad at coglit.ecs.soton.ac.uk (Stevan Harnad) Date: Sun, 17 Oct 1999 14:12:56 +0100 (BST) Subject: PSYC Call for Book Reviewers: Cognitive Mapping/Golledge Message-ID: PSYCOLOQUY CALL FOR BOOK REVIEWERS Below is the Precis of "Wayfinding Behavior: Cognitive mapping and other spatial processes" (685 lines). This book has been selected for multiple review in PSYCOLOQUY. If you wish to submit a formal book review please write to psyc at pucc.princeton.edu indicating what expertise you would bring to bear on reviewing the book if you were selected to review it. (If you have never reviewed for PSYCOLOQUY or Behavioral & Brain Sciences before, it would be helpful if you could also append a copy of your CV to your inquiry.) If you are selected as one of the reviewers and do not have a copy of the book, you will be sent a copy of the book directly by the publisher (please let us know if you have a copy already). Reviews may also be submitted without invitation, but all reviews will be refereed. The author will reply to all accepted reviews. Full Psycoloquy book review instructions at: http://www.princeton.edu/~harnad/psyc.html http://www.cogsci.soton.ac.uk/psycoloquy/ Relevant excerpts: Psycoloquy reviews are of the book not the Precis. Length should be about 200 lines [c. 1800 words], with a short abstract (about 50 words), an indexable title, and reviewer's full name and institutional address, email and Home Page URL. All references that are electronically accessible should also have URLs. ----------------------------------------------------------------------- psyc.99.10.036.cognitive-mapping.1.golledge Sun Oct 17 1999 ISSN 1055-0143 (32 paragraphs, 14 references, 685 lines) PSYCOLOQUY is sponsored by the American Psychological Association (APA) Copyright 1999 Reginald Golledge WAYFINDING BEHAVIOR: COGNITIVE MAPPING AND OTHER SPATIAL PROCESSES. [John Hopkins University Press, 1999 xviii, 428pp, ISBN: 0-8018-5993-X] Precis of Golledge on Cognitive-Mapping Reginald G. Golledge Department of Geography University of California Santa Barbara Santa Barbara CA 93106-4060 U.S.A. golledge at geog.ucsb.edu ABSTRACT: This is an edited volume of essays by psychologists, biologists, cognitive scientists, computer scientists, and geographers on wayfinding by humans and other species. It addresses the extent to which cognitive maps may be universal, and produces evidence that humans, apes, some birds and some small mammals appear to behave as if they have internal representations that guide wayfinding processes in a map-like manner. Evidence also shows that insects, some mammals, and perhaps some birds may not evince such guided behavior, but rely more on spatial updating by dead-reckoning or pilotage. The multiple disciplinary views of wayfinding and navigation by humans and other animals gives the volume a distinctly different content from other available books. KEYWORDS: cognitive map; internal representation; navigation; navigation; path integration; place cells; wayfinding. I. INTRODUCTION 1. The use of maps in forms ranging from dirt drawings to stone carvings, from rice paper scrolls to Automobile Association trip-tiks, from topographic map sheets to disposable tactile strip maps appears to be a cultural universal (Uttal 1997). Maps both record what is known and remembered about an environment and act as wayfinding aids. In the absence of these artifacts, humans and other animals rely on internal representations or stored memories of experienced environments. It is frequently assumed that these stored memories, now commonly referred to as cognitive maps or internal spatial representations, are used to guide travel. Cognitive maps are always there: they cannot be left at home, torn to pieces by fractious children, rendered apart and pieced together incorrectly, so that map reading errors result in a traveler becoming lost. But they do have their problems. 2. The idea that animals also possess internal spatial representations resulted in Tolman's first identification of the term cognitive map. Rather than state that animals, particularly the rats that took short cuts through his mazes stored spatial information as a map, Tolman used the term metaphorically. In other words, he suggested that the animals used in his studies appeared to be able to use spatial information as though the places they remembered were recorded in a maplike manner. For decades, controversy has raged over whether animals do have cognitive maps or if they have other forms of internal spatial representations that allow them to behave as if they were being guided by a map-reading operation. After decades of research in zoology, other biological sciences, and experimental psychology, in particular, various alternatives have been posed to account for successful animal travel behavior. Many have argued that the practice of returning directly to home after a meandering search for food by many nonhuman species indicates that the species did continuous spatial updating, then returned home by a procedure well known to ocean shipping or aircraft pilots the process of dead reckoning. Called path integration, this process enables a traveler to constantly update their current position with respect to an origin without recording details of the path already followed. Because there is no need for a memory trace of the path, route retrace may be difficult or impossible. The need of many foragers who are partly responsible for feeding other members of their species to return home with food, appears to make the short cut (or 'beeline' or 'crow-fly') return trip the more reasonable option. If food is consumed at the spot on which it is found, safety considerations might dictate an immediate shortest-distance return. 3. Flying insects and avian species appear to use landmarks, sun compasses, magnetic compasses, or other celestial guides to help them with migratory and shorter-distance travel. Naturally enough, questions have arisen as to whether the landmarks used are captured as a perspectively viewed retinal image, or whether their configuration or layout is either stored and recalled in sequence as a route is followed, or represented as layouts or configurations similar to a survey (or overview-based) representation of a large-scale and complex space. 4. Despite the existence of these two vigorous research areas, focused on nonhuman and human travel respectively, until recently there have been few deliberate attempts to combine the two literatures. This lack of attention provided the rationale for a small seminar funded by the Borchard Foundation at the Chateau de la Bretesche in July 1996. The purpose of the meeting was to bring together researchers from both the human and nonhuman research domains who had specialized in navigation or wayfinding behavior and who were familiar with the idea of cognitive mapping and the potential role that cognitive maps might play in wayfinding behaviors. As the guests of the foundation's director, William F. Behling and his wife, Betty, at the Chateau de la Bretesche, nine contributors to this volume first presented position statements on the relationships between cognitive maps and wayfinding in humans and other species. To supplement this group's expertise, other scientists were invited to add chapters to the book. 5. The term 'cognitive maps' is used throughout this book to refer to the internal spatial representation of environmental information. Its use varies, from the metaphorical ('as if' the information was stored in maplike format), to a hypothetical construct. 6. The term 'spatial representation' is also used throughout the book. This might be regarded as a shorthand notation for the organization of components of spatial knowledge or other partly investigative processes (e.g., neurophysiological structures and place cells, cell assemblies, phase sequences). The term also can be used metaphorically, involving an 'as if' quality, particularly when referring to purported maplike properties of representations. It has also been used as an intervening variable in which it is interpreted as a 'note' attached to an economical grouping of measured variables in a statement of functional relations between other measured variable. 7. Structurally, this book is divided into four sections, ranging from wayfinding and cognitive mapping in humans operating in different scenarios, to examinations of special human navigation processes (e.g. without sight), to studies of wayfinding by various non-human species, and the neurobiological bases of environmental knowledge acquisition and use. Each section is now summarized in turn. II. SECTION I: HUMAN COGNITIVE MAPS AND WAYFINDING 8. This first section explores the strong theoretical and empirical links between cognitive maps (or the internal representation of environmental information); the cognitive mapping process itself; the internal manipulation of information in the form of spatial choice and decision making, and the directed acts of human wayfinding through simple or complex environments. The evidence is clear and overwhelming that human wayfinding is directed and motivated, and follows sets of procedural rules whose content and structure are the focus of much ongoing research. The consensus is clear: humans acquire, code, store, decode, and use cognitive information as part of their navigation and wayfinding activities. Although over the centuries they have developed numerous ways of supplementing personally stored environmental information (e.g., maps, written descriptions, and various forms of image representations), it appears that humans rely on personal cognitions to make many spatial decisions, and to guide their movement behavior. There is evidence that internal representations and their externalizations (spatial products) do not necessarily match well, and that the existence of fragmented, incomplete, or distorted cognitive maps appears to account for many behaviors that might otherwise be labeled as spatially irrational. 9. The purpose of this part of the book is to examine sets of concepts deemed relevant to both human wayfinding and cognitive mapping. There are two chapters by geographers (Golledge, and Stern and Portugali), and two by psychologists (Allen and Grling). Although some disciplinary perspectives are evident, there is much overlap and common concern. The first two chapters, by Golledge and Allen (respectively), provide overviews and summaries of theories and concepts relevant throughout the entire book. The following chapters by Grling, and Stern and Portugali have a tighter focus: Grling emphasizes the sequential spatial choice processes so important to human wayfinding, and Stern and Portugali emphasize decision making in urban environments, the complex scenarios in which most humans live and interact. All four chapters contain examples of relevant research. 10. In the first chapter, Golledge reviews critical definitions relating to cognitive maps and wayfinding. He provides an overview of the role of cognitive mapping in human wayfinding and describes the processes of acquiring and storing spatial information about large-scale complex environments. Further, he discusses how humans record and represent environmental knowledge. The role played by landmarks and routes in anchoring knowledge and in wayfinding is examined, and the differences between path following and route-based environmental learning are explored. Errors commonly related to encoding, decoding and internally manipulating cognized spatial data are highlighted. Wayfinding by humans in contexts other than with landmark usage is also examined, and an elaboration of errors commonly found in human wayfinding follows. Throughout, ties are made to treatments of similar problems in later chapters that focus on the nonhuman domains of internal spatial representations and wayfinding. 11. In the second chapter, Allen provides insights into the nature of spatial abilities and the role they play in cognitive mapping and wayfinding procedures. He places emphasis on the concept of individual differences in spatial cognition and in behavior. Allen argues that the scientific literature in psychology and geography contains a vast number of studies concerned with spatial abilities and a growing body of research on wayfinding, although little has been done to establish the relevance of the former for the latter. Thus the question of why some individuals are better than others at wayfinding has been difficult to address. Allen suggests that a potentially informative way to think of wayfinding is to differentiate between wayfinding tasks and wayfinding means. Tasks include traveling to a previously known destination, exploration with the purpose of returning home, and traveling to a novel destination. Means include oriented search, following a continuously marked trail, piloting (between landmarks), habitual locomotion, path integration, and reference to a cognitive map. Spatial abilities in the past have been examined from psychometric, information processing, developmental, and neuropsychological perspectives. Allen suggests that broad fami1ies of abilities involved in the identification of manipulable objects, those involved in anticipating the trajectory and speed of moving objects, and those involved in supporting oriented travel within large-scale environments summarize the dominant research themes. He implies there is considerable utility associated with the concept of interactive common resources for cognitive and perceptual-motor tasks. The result of the use of spatial abilities is support-oriented travel, but they also serve as a resource for acquiring additional environmental knowledge. Cognitive maps are considered as knowledge of places and cognitive mapping includes rules for establishing spatial relations among such places. 12. Next, Grling discusses human information processing in sequential spatial choice, which summarizes the essential acts involved in wayfinding. He begins with the premise that human locomotion in space is goal-directed. Spatial orientation and navigation are, therefore, primarily means of monitoring travel plans. Travel plans are developed prior to initiating movement. The chapter focuses on the formation of travel plans and their consequent execution. Such planning entails spatial choices that are multiattribute, sequential, and stated. He summarizes research on how people process information when solving the traveling salesman problem (i.e., finding the shortest distance between an origin and a set of destinations that might be sequentially visited). He details research on how time and priority are traded off against spatial attributes in sequential spatial choices. 13. Stern and Portugali next examine the relationship between environmental cognition and decision making in urban navigation. They define urban navigation as a sequential process of decision making concerning route choice. They claim that traditionally many choice situations are described by a 'black-box approach', which does not specify choice rules but rather deals only with the relationship between input and output variables. In most of these models a cognitive explanatory mechanism of the choice process is missing. Their chapter presents two complementary conceptual frameworks as possible ways to solve this problem. The first is the inter-representational network (IRN), and the second is decision field theory (DFT). It is suggested that both frameworks can explain the dynamics and high variability in the choices of persons navigating in urban environments. 14. In the exploration of human wayfinding and its various components as illustrated in this first section, the importance of individual differences, those between males and females, and variations according to one's spatial abilities are reviewed. III. SECTION II: PERCEPTUAL AND COGNITIVE PROCESSING OF ENVIRONMENTAL INFORMATION 15. In this part, three chapters explore cognitive processes and human navigation in a variety of contexts, including an extensive investigation of path integration by humans covering wayfinding without vision; updating an object's orientation and location during nonvisual navigation; exploring the geometrical constraints and calibration of action-representation couplings, and relating perceptual processes to various navigation requirements. Focusing primarily on aspects of human perception and cognition with respect to wayfinding, these authors explore nontraditional domains to show the versatility of relevant theories and concepts. Although vision is accepted as the most important spatial sense, there is no doubt that blind or vision-impaired humans can become competent independent travelers using simple cognitive processes and aids such as the white cane, guide dog, or a variety of recently developed auditory navigational aids. 16. In the first of these chapters (chapter 5) Loomis, Golledge, Klatzky, and Philbeck discuss the process of human navigation by path integration, a process that until recently was recognized more in the nonhuman domain. They begin by clearly defining two types of processes influential in wayfinding piloting and path integration. The recent literature is replete with misconceptions of the nature of these processes, but little is left in doubt following their clear and comprehensive discussion. Navigation by humans, animals, and machines is accomplished using two distinct methods. Piloting is the determination of current position and orientation using landmark information in conjunction with a map, either external or internal. Path integration is the updating of position and orientation on the basis of velocity and acceleration information about self-movement. The chapter begins with a consideration of a number of models of path integration. Following is a review of the empirical research on human path integration, with a focus on controlled experimental investigations. Such investigations have been carried out using two distinct tasks: return-to-origin after the passively guided traverse of an outbound path, and perceptually directed action, whereby the person sees or hears a target and then, with the target extinguished, attempts to indicate its position by actively locomoting toward it or by pointing in its direction during locomotion that passes by the target. 17. In Chapter 6, Amorim discusses a neurocognitive approach to human navigation. He suggests that human navigation is viewed as a result of the interplay of neurocognitive functions. Spatial updating and frames of reference constitute the two concepts of maximum interest in this work. He provides experimental evidence on the role of reference frames in computing locations in space, as well as on the effect of two processing modes for the updating of an object's location and orientation. Amorim uses an information-processing approach (commonly used in cognitive psychology) in an effort to understand human processes of updating an object's location and orienting it with respect to a bounding frame of reference. To localize a person in the environment as well as localize an object the environment contains, Amorim suggests that the acquisition, coding, and integration of sensory information (both perceptual and representational) are necessary. Building on the model of visuo-spatial cognition proposed by Kosslyn (1991), Amorim offers two studies; one investigates the role of reference frames in computing locations in space, whereas the other compares two processing modes for the updating of an object's location and orientation in space. In interpreting the results of these experiments he evaluates the neurocognitive approach to the study of the pathological causes of topographical disorientation. 18. In chapter 7, Rieser examines action-representation couplings, focusing in particular on the geometrical constraints on such calibrations. He argues that perception and action are coupled, so that motoric actions result in dependable changes in the actor's perspective. For example, during locomotion the structure of an actor's perspective visibly rotates and translates in directions and at rates that fit with the geometry and rate of locomotion. This coupling provides a chance for perceptual-motor learning. While walking with vision, people learn the covariation of optical (and possibly nonoptical) flow and afferent-efferent input associated with the biomechanical activities of walking. This learning, in turn, provides the basis for the coupling of representation and action. Representation is coupled with action in working memory in analogous ways. When acting without vision, people are knowledgeable about the resulting changes in their perspective. So for example, after viewing their surroundings and then walking without vision, people are able to keep up to date on the changing self-to-object distances and directions relative to their remembered surroundings. IV. SECTION III: WAYFINDING AND COGNITIVE MAPS IN NONHUMAN SPECIES 19. In previous sections we focused on humans, in whom cognitive processing is well established, but the tie to navigation and wayfinding is not strongly defined. In this section the authors focus on navigation and wayfinding by nonhuman species, in which the presence of cognitive maps is being strongly debated. In these chapters, biological and ecological scientists examine wayfinding and discuss the possibility that different species have and use cognitive maps. 20. Etienne, Maurer, Georgakopoulos, and Griffin begin (in chapter 8) with a review of the significance of dead reckoning or path integration and landmark use in the representation of space. In many ways this provides a view that complements chapter 5 by Loomis et al., which presents a human navigator's view of the same process. In particular they examine suggestions that dead reckoning (which does not involve learning an environment) seems more dominant in nonhumans, whereas landmark-guided movement may be more dominant in humans. The problem of how different species combine the systems in wayfinding is examined in great detail. Drawing on examples from their group's work with small mammals, Etienne et al. suggest that animals may well have a simple cognitive map that helps their memory for routes and places (such as sources of food or food storage areas). 21. But not all animals may have such cognitive maps. In this chapter, Etienne et al. begin from the viewpoint that spatial representation as defined originally by Tolman (1948) and more recently by O'Keefe and Nadel (1978) refers to a high level of spatial information processing. They use the term cognitive map to imply that a subject organizes the familiar environment as a system of interconnected places and that it applies a set of transformation rules to this system, which may consist of a limited number of complementary operations (such as those hypothesized by Piaget 1937), or that optimize goal-directed movements. Thus whether human or nonhuman, a subject must be able to pilot and perform new route selection before being credited with possessing a cognitive map. 22. The authors define piloting in terms of planning and performing a goal-directed path by deducing an itinerary from the memorized spatial relations between a goal and a traveler's current position, while new route performance implies an ability to select the most economical alternative path (including shortest path and shortcuts) in both familiar and unfamiliar settings. If a cognitive map alone is used, then piloting and path following must take place without either the use of beacons or reference to external landmarks. Etienne et al. argue that the general literature has yet to yield convincing evidence that spatial knowledge reaches this degree of coherence in species other than primates. They suggest using the term spatial representation, or more precisely, the representation of locomotive space, for their work with nonprimate animal species. Thus their chapter directly addresses the question of the universality of cognitive maps by suggesting that whereas spatial representation may be universal, cognitive maps may develop only in a limited number of species. They then point out that the attribution of specific systems of representation to different species poses severe problems. They argue that if one ascribes to an animal or a young child particular forms of spatial representation, inevitably one begins by analyzing subjects' behavior in specific functional contexts to see how observed behaviors fit certain aspects of the environment. 23. The authors make a strong statement that all sedentary species adapt their locomotor behavior to relevant features in the spatial environment in order to reach their goals without getting lost. Thus the observed correspondence between behavior and functionally meaningful aspects of the environment gives insights into what the traveler knows about the environment and thus how the external world is represented or modeled. The authors then examine the process of dead reckoning, with and without the possible use of ancillary landmarks. They report that many theories of navigation emphasize that dead reckoning (path integration) plays a significant role in spatial representation and wayfinding across the entire animal kingdom from insects and other invertebrates to mammals (Gallistel 1990). Then, building on this fascinating introduction, the authors examine the role of dead reckoning in the representation of space in a comparative perspective, including hymenopterans and rodents. They describe how insects and mammals use dead reckoning as current route-based information and how they use landmark-place associations as long-term location-based references. They then consider how the species previously mentioned represents space on the basis of route-based and location-based information, and on the interaction between these two categories of references. 24. In chapter 9, Judd, Dale, and Collett examine the fine structure of view-based navigation in insects. They begin by asserting that insects learn landmarks as two-dimensional views. These views are highly dependent on vantage points, so that even over a relatively short section of a foraging trip, the insect's view of a nearby landmark will change appreciably. Insects simplify the problems of using such retinotopic views for navigation in a number of ways. For example, bees and wasps restrict the range of directions in which they approach a familiar place so that they capture roughly the same sequence of retinal images from visit to visit (i.e.. approach from the same perspective view). They are guided into the vicinity of the goal by aiming at a nearby beacon landmark. Because of changes in image size and shape, a single stored view of the beacon is unlikely to allow the insect to recognize it over the whole range of possible approaches. In addition, the authors claim that wood ants are shown to take several 'snapshots' of a beacon at different distances in the early stages of learning a new environment. Once close to a beacon, the insect relinquishes fixation either to approach another beacon or to approach the goal. This transition is achieved by linking a stereotyped action to a frontally stored view of the beacon. By this means the insect can acquire a standard view of the next beacon or arrive at a point close enough to the goal to allow image matching of the goal itself or the nearest landmark. The goal is then pinpointed by moving so that the image on the retina matches the view of nearby landmarks. The authors go on to suggest that there is surprising similarity in the motor constraints and landmark strategies of real insects and those of simple simulated 'creatures' they have 'evolved' artificially. Again, the parallel between human and non-human species stands out. 25. Moving from ground-based animals and low-flying insects to birds, the internationally acclaimed team of Wolfgang and Roswitha Wiltschko (chapter 10) discuss compass orientation and basic elements in avian orientation and navigation. Birds face orientation tasks in two behavioral contexts: homing and migration. Because of the long distances involved in migration, birds must establish contact to their goal indirectly via an external reference. Three such mechanisms have been described: a magnetic compass based on the field lines of the geomagnetic field and two compass mechanisms based on celestial cues, namely a sun compass and a star compass. To use a compass, birds must first determine the compass course leading to their destination. For homing, experimental evidence indicates that experienced pigeons can derive the home course from site-specific information obtained at the starting point of the return flight. Their ability to do this even at distant, unfamiliar sites has led to the concept of the navigational 'map', which is a directionally oriented representation of the distribution of environmental gradients within the home region. It can be extrapolated beyond the range of direct experience. Birds determine their home course by comparing local values of these gradients with the home values. The 'map' is based on individual experience. 26. During an early phase in life, young pigeons derive their home course from directional information collected during the outward journey. On spontaneous flights, they record prominent landmarks and changes in navigational factors and combine this information with the direction flown to form the navigational map. Once the map is established, it is preferentially used, because it permits the correction of errors. The navigational map is a cognitive map because it allows novel routes; it differs from cognitive maps discussed for other animals by the size of the area covered and by including continuous factors like gradients. In migration, birds must reach a distant region of the world. The course leading to this goal area is constant; the birds possess genetically coded information on their migratory direction. The conversion of this information into an actual compass course requires external references, which are provided by celestial rotation and by the geomagnetic field. Celestial rotation indicates a reference direction away from the celestial pole, whereas the magnetic field defines a specific deviation from this course, resulting in the population-specific migration course. Both types of cue continue to interact during migratory flights. Depending on the nature of the orientation tasks, birds make use of innate information or of individual learning processes. In both strategies, however, external references provided by compass mechanisms are essential components. 27. Thinus-Blanc and Gaunet (chapter 11) discuss the cognitive map as an internal representation of an environment where places and their spatial relationships (such as angles and distances) are charted. This notion has been extensively criticized in the past by Thinus-Blanc because the expression is antinomic and can easily lead to misunderstanding. The authors point out that 'cognitive' refers to dynamic processes and 'map' refers to a static picture of the real world. To this extent, the term cognitive mapping is functionally more correct. Internal spatial representations are said to be useful for orienting in a given environment just as they contribute to the organization of new spatial information as it is accrued. Thus Thinus-Blanc and Gaunet argue that spatial representation may be viewed as maps of the environment but more appropriately should be viewed as cognitive or active information seeking structures. They draw on data from animal and human studies and related theoretical work to support this hypothesis. V. SECTION IV: THE NEURAL AND COMPUTATIONAL BASES OF WAYFINDING AND COGNITIVE MAPS 28. In this part cognitive neuroscientists Nadel (University of Arizona) and Berthoz (Laboratoire de Physiologie de la Perception et de l'Action, College de France) respectively examine the neural bases of wayfinding and cognitive maps, and computer scientist Chown (Bowdoin College) discusses their implications for computation and artificial intelligence-based travel. 29. In chapter 12, Nadel provides an overview of the neural mechanisms of spatial orientation and wayfinding. He suggests that work on the neural bases of wayfinding in mammals has intensified in recent years, building on the discovery of place neurons in the hippocampus. The cognitive map theory of hippocampal functioning, first put forward by O'Keefe and Nadel in 1978, suggests that this brain structure is the core of an extensive neural system subserving the representation and use of information about the spatial environment. Nadel argues that evidence supporting this theory comes primarily from brain lesion and neurophysiological recording studies. The former showed that damage in the hippocampus system invariably impairs the ability of animals and humans to learn about, remember, and navigate through environments, while the latter show that neurons in this system code for location, direction, and distance, thereby providing the elements needed for a mapping system. Current work in this area focuses on which stimuli control the activity of these neural elements, and how the system is used in behavior. He cites the fact that the roles of external and internal sources of information are under active investigation. 30. In the next chapter, Berthoz and his associates examine the neural bias of spatial memory during locomotion. This chapter addresses the question of the mechanisms that underlie the capacity to memorize routes and to use this spatial memory for guiding and steering of locomotion. A review is presented of several paradigms used in their laboratory to study this question. First, some previous studies, which have shown that vestibular information about head rotation and translation can be used by the brain to estimate distances are reviewed. Berthoz et al. claim that such use has been shown by the vestibular memory contingent 'saccade task', both in normal subjects and in neurological patients. Second, some recent experiments that use the task of walking along a triangular path with or without vision are described. During this task, head position and velocity are measured by video-computerized techniques. Two main results have been obtained: (1) They have discovered that the head anticipates the body movement during walks around a corner: this anticipation also exists in darkness, suggesting that the orienting system is driven by an internal representation of the trajectory and that the brain uses a strategy of guiding locomotion by gaze (go where you look) even in darkness. (2) When vestibular-deficient patients perform the task, they seem to control the total distance but not the direction, suggesting a dissociation between the control of distance and direction in this locomotor pointing task. They also describe a second paradigm of circular locomotion, during which subjects were asked to walk around a circular path with or without vision. Here again, the measure of the kinematics parameters of head movement indicates both an anticipation of head direction and a dissociation between the control of distance and direction, and provides clarification of the frequently misinterpreted concepts of course and heading. Finally, they review a number of recent results which may lead to an understanding of the neural basis of both anticipation and the role of vestibular cues in the steering of locomotion. 31. In the final chapter, Chown discusses error tolerance and generalization in cognitive maps. He asserts that human cognitive maps are not precise, complete, nor necessarily accurate. Because navigation is so important in everyday life, it is not easy to understand why humans have evolved an internal representation of space that appears to have such basic flaws. The theme of this chapter is that it is exactly the sketchy nature of human cognitive maps that make them such a powerful tool for navigation. There is growing evidence from artificial intelligence and robotics that in real environments, useful representations cannot be achieved without sacrificing completeness and precision. Further, it can be shown that the sketchy nature of cognitive maps more naturally lends itself to error tolerance and generalization than would be the case with alternative structures. Cognitive maps may be sketchy but the information they do store is usually sufficient for human needs. The relationship between human needs and how cognitive maps encode information is discussed in a proposed model called PLAN. VI. FINAL COMMENT: 32. The more we know about how humans or other species can navigate, wayfind, sense, and record and use spatial information, the more effective will be the building of future guidance systems, and the more natural it will be for humans to understand and control those systems. The question of which of the many cognitive mapping, navigational, or wayfinding procedures and behaviors should be taken as the role model for future systems remains unanswered at this stage. Knowing the advantages and disadvantages, the strengths and the shortcomings, the idiosyncrasies and the universals of spatial knowledge acquisition and storage and wayfinding behavior can only lead to the development of systems that are as endemic as path integration, as powerful as cognitive mapping, and as anchored as landmark usage, and that possess the versatility to handle both view-centered and object-centered modes of recording or experiencing new environments. REFERENCES: Gallistel, C. R. (1990). The organization of learning. Cambridge, MA: MIT Press. Kosslyn, S. M. (1991). 'A cognitive neuroscience of visual cognition: Further developments.' In R. H. Logie & M. Denis (Eds.) Mental Images in Human Cognition (pp.351-381). Amsterdam: Elsevier Science Publishers. O'Keefe, J., & Nadel, L. (1978). The hippocampus as a cognitive map. Oxford: Oxford University Press. Piaget, J. (1937). La construction du reel chez l'enfant. Paris: Delachaux et Niestl, Neuchtel. Tolman, E. C. (1948). 'Cognitive maps in rats and men.' Psychological Review, 55, 189-208. Uttal, D.H. (1997, April) 'Seeing the big picture: Children's mental representation of spatial information acquired from maps.' Paper presented at the 93rd annual meeting of The Association of American Geographers, Ft. Worth, TX. List of author names with chapter titles and page numbers: 1: Human Wayfinding and Cognitive Maps, 5-46. Reginald G. Golledge 2: Spatial Abilities, Cognitive Maps, and Wayfinding: Bases for Individual Differences in Spatial Cognition and Behavior, 46-81. Gary L. Allen 3: Human Information Processing in Sequential Spatial Choice, 81-99. Tommy Grling 4: Environmental Cognition Decision Makaing in Urban Navigation, 99-121. Eliahu Stern Juval Portugali PART II PERCEPTUAL AND COGNITIVE PROCESSING OF ENVIRONMENTAL INFORMATION 5: Human Navigation by Path Integration, 125-152. Jack M. Loomis Roberta L. Klatzky Reginald G. Golledge John W. Philbeck 6: A Neurocognitive Approach to Human Navigation, 152-168. Michel-Ange Amorim 7: Dynamic Spatial Orientationa and the Coupling of Representation and Action, 168-191. John J. Rieser PART III WAYFINDING AND COGNITIVE MAPS IN NONHUMAN SPECIES 8: Dead Reckoning (Path Integration), Landmarks, and Representation of Space in a Comparative Perspective, 197-229. Ariane S. Etienne Roland Maurer Josephine Georgakopoulos Andrea Griffin 9: On the Fine Structure of View-Based Navigation in Insects, 229-259. Simon P. D. Judd Kyran Dale Thomas S. Collett 10: Compass Orientation as a Basic Element in Avian Orientation and Navigation, 259-294. Roswitha Wiltschko, Wolfgang Wiltschko 11: Spatial Processing in Animals and Humans: The Organizing Function of Representations for Information Gathering, 294-309. Catherine Thinus-Blanc Florence Gaunet PART IV THE NEURAL AND COMPUTATIONAL BASES OF WAYFINDING AND COGNITIVE MAPS 12: Neural Mechanisms of Spatial Orientation and Wayfinding: An Overview, 313-328. Lynn Nadel 13: Dissociation between Distance and Direction during Locomotor Navigation, 328-349. Alain Berthz Michel-Ange Amorim 14: Error Tolerance and Generalization in Cognitive Maps: Performance without Precision 349. Eric Chown REFERENCES 371 CONTRIBUTORS 415 INDEX 419 From curt at cassandra.ucr.edu Mon Oct 18 18:45:22 1999 From: curt at cassandra.ucr.edu (Dr. Curt Burgess) Date: Mon, 18 Oct 1999 15:45:22 -0700 Subject: quant & developmental cognition position at Univ Calif, Riverside Message-ID: Just wanted to make sure that any interested parties on the Connectionist email distribution list knew about this position. Curt Burgess ---------------------- The Department of Psychology, University of California, Riverside, invites applications for a tenure-track Assistant Professor position in Developmental Psychology, beginning July 1, 2000. The Ph.D. degree is required at the time of the appointment. Applicants in all areas of developmental psychology are encouraged to apply. Preferred applicants are those with an interest in cognitive development with strong quantitative skills, who can contribute to the graduate program in quantitative psychology. We are seeking a developmental psychologist with a strong program of substantive research that complements that of the current developmental faculty, whose emphases include cognitive and social development, and processes of socialization. Applicants should also be committed to excellence in undergraduate and graduate education. The deadline for review of completed applications begins January 10, 2000 and continues until the position is filled. Interested candidates should send their curriculum vitae, reprints if available, a cover letter describing research and teaching interests, and arrange to have three letters of recommendation sent to: Chair, Developmental Psychology Search Committee Box I-CN Department of Psychology University of California - Riverside Riverside, CA 92521 The Riverside campus of the University of California is growing rapidly and has an excellent psychology department with a strong record of success in research, teaching, and extramural funding. For information on the Department of Psychology, see our web site at: http://www.psych.ucr.edu/. The campus is centrally located in Southern California, about 50 miles east of Los Angeles and less than an hour's drive from the area's mountains, deserts, and beaches. The University of California, Riverside is an equal opportunity employer/affirmative action employer. --------------------- Dr. Curt Burgess, Computational Cognition Lab Chair of Graduate Admissions for Psychology Department of Psychology, University of California 1419 Life Science Building Riverside, CA 92521-0426 URL: http://locutus.ucr.edu/ Internet: curt at cassandra.ucr.edu MaBellNet: (909) 787-2392 FAX: (909) 787-3985 From wolpert at hera.ucl.ac.uk Tue Oct 19 06:48:25 1999 From: wolpert at hera.ucl.ac.uk (daniel wolpert) Date: Tue, 19 Oct 1999 11:48:25 +0100 (GMT) Subject: Positions in Computational Sensorimotor Control Message-ID: <14348.19526.664653.987081@perseus.ion.ucl.ac.uk> ********************************************************************* Nature advertisement October 21st Institute of Neurology, University College London Applications are invited for three posts to work with Dr. Daniel Wolpert under a Wellcome Programme Grant entitled Computational Human Sensorimotor Control. These positions are available for three years in the first instance with a starting date from January 2000, but candidates who wish to start later in 2000 are also encouraged to apply. Further details of all posts are available on http://www.hera.ucl.ac.uk TWO POSTDOCTORAL RESEARCH FELLOWS Two Fellows are required with PhDs relevant to the study of human motor control. One Fellow will work on modelling of the motor system using optimal control and connectionist approaches and should have a background in a computational field (e.g. Computational Neuroscience, Engineering, Physics, Maths). One Fellow will work on psychophysical studies of human motor control and should have a background in an experimental field (e.g. Neuroscience, Psychology). RESEARCH ASSISTANT The Research Assistant will provide software and hardware support for the group. Experience with C/C++ programming essential and experience with NT, Linux, OpenGL, external device interface/control and electronics would be advantageous. Starting salary for all positions is up to ?27,586 pa inclusive, `depending on experience. Applications (2 copies of CV and names of 3 referees) to Miss E Bertram, Assistant Secretary (Personnel), Institute of Neurology, Queen Square, London WC1N 3BG (fax: +44 (0) 171 278 5069) by 1 December 1999. Informal enquiries to Dr. Daniel Wolpert (email: wolpert at hera.ucl.ac.uk; tel: 0171 837 3611 ext 4183; fax: 0171 813 3107). Working toward Equal Opportunity ********************************************************************* From Richard.Woesler at t-online.de Wed Oct 20 10:00:16 1999 From: Richard.Woesler at t-online.de (Richard Woesler) Date: Wed, 20 Oct 1999 16:00:16 +0200 Subject: Thesis Announcement Message-ID: -------------------------------------------------------------------------------- Dear Connectionists, I am pleased to announce the availability of my thesis. Regards, Richard Woesler -------------------------------------------------------------------------------- Title: Neural Networks for Object Segmentation in Amphibians and an algorithmic Classification of Neurons ISBN 3-89722-073-3, Logos-Verlag Berlin, Germany 1999 Advisors: Prof. Helmut Schwegler, head of the Institute of Theoretical Neurophysics at the University of Bremen, and Prof. Gerhard Roth, head of the Institute of Brain Research at the University of Bremen, Germany Language: german Abstract: Neural network models for object segmentation in amphibians, especially in the tongue projecting salamander Hydromantes italicus, constitute the main part of the text. Firstly, we carry out a simple double-dummy experiment in which the animal has to choose one dummy out of two identical, coherently moving prey-dummies. Hydromantes is able to select one of them and to snap at it with high accuracy. The question arises how this object-selection process, the segmentation of the selected object from the rest of the visual scene, and the high accuracy are realized within the amphibian brain. First, we present a simplified model, our Base Model. This model detects moving contrast edges, binds these edges to objects, and selects one object. We derive analytical results concerning the performance level of the network. The results are similar to those known from amphibians. Second, we extend this model to our Amphibian Model where a number of recent biological data is taken into account, especially data about the retina and the tectum opticum which is an important brain region with respect to prey-capture behaviour and which is homologous to the mammalian superior colliculus, and data about the nucleus isthmi which is a relatively small brain region connected reciprocally with the tectum opticum and which is homologous to the mammalian nucleus parabigeminalis. This model is a complete neural network for object segmentation from retinal photoreceptors to the segmentation of the selected object. The model yields possible explanations for many biological findings: e. g., for the width distribution of the receptive fields of neurons of the tectum opticum, or for the range of velocities of objects for which Hydromantes italicus shows prey-capture behaviour. The model contains a spotlight network which enhances the retinotectal transmission in a certain spotlight region to further neurons, which we call the multiplicative coarse coding neurons. The spotlight network should segment the selected object from the rest of the visual scene, and the multiplicative coarse coding neurons should encode its place with high accuracy. We describe in detail how this network can be realized in amphibians. Third, we present the computer model Simulander III with which we simulate the described double-dummy experiment. Indeed the model shows the high object-selection ability and the high accuracy known from Hydromantes. We give an outlook to various model extensions, especially to a model of recognition of the absolute width of the selected object and of recognition of a stepwise prey movement. These abilities are also known from Hydromantes. Additionally, we discuss a model extension which includes a neural network of synchronizing neurons. In the second part of the text, we consider the classification of neurons of the tectum opticum of Hydromantes italicus according to their responses to different prey-dummies. We are particularly interested in the question whether the response types form distinct classes or a continuum. We classify the data with an algorithm that we have developed using known classification methods. The classification result suggests that the data form a continuum with some accumulations. Finally, we discuss the possible biological reasons for a continuum and possible advantages of a continuum. -------------------------------------------------------------------------------- ******************************************** Dipl.-Phys. Dr. Richard Woesler August-Schlueter-Str. 39 48249 Duelmen Germany Phone +49 (0) 2594-991182 Fax +49 (0) 2594-948752 http://home.t-online.de/home/Richard.Woesler ******************************************** From giles at research.nj.nec.com Wed Oct 20 19:29:34 1999 From: giles at research.nj.nec.com (Lee Giles) Date: Wed, 20 Oct 1999 19:29:34 -0400 (EDT) Subject: call for participation - NIPS99 Message-ID: <199910202329.TAA10434@alta.nj.nec.com> CALL FOR PARTICIPATION -- NIPS*99 http://www.cs.cmu.edu/Web/Groups/NIPS Neural Information Processing Systems -- Natural and Synthetic Monday November 29 - Saturday December 4, 1999 Denver, Colorado This is the thirteenth meeting of an interdisciplinary conference which brings together cognitive scientists, computer scientists, engineers, neuroscientists, physicists, and mathematicians interested in all aspects of neural processing and computation. The conference will include invited talks as well as oral and poster presentations of refereed papers. The conference is single track and is highly selective. Preceding the main session, there will be one day of tutorial presentations (November 29), and following it, there will be two days of focused workshops on topical issues at a Breckenridge (December 3-4). Major categories of accepted papers include the following topics: Algorithms and Architectures: supervised and unsupervised learning algorithms, model selection algorithms, active learning algorithms, feedforward and recurrent network architectures, localized basis functions, mixture models, belief networks, graphical models, Gaussian processes, factor analysis, topographic maps, combinatorial optimization, hybrid symbolic-subsymbolic systems. Applications: handwriting recognition, sequence analysis, expert systems, fault diagnosis, medical diagnosis, analysis of medical images, data analysis, database mining, information retrieval, network traffic, music processing, time-series prediction, financial analysis. Cognitive Science/Artificial Intelligence: perception and psychophysics, neuropsychology, cognitive neuroscience, development, conditioning, human learning and memory, attention, language, natural language, reasoning, spatial cognition, emotional cognition, conceptual representation, neurophilosophy, problem solving and planning. Implementations: analog and digital VLSI, optical neurocomputing systems, novel neurodevices, computational sensors and actuators, simulation tools. Neuroscience: neural encoding, spiking neurons, synchronicity, sensory processing, systems neurophysiology, neuronal development, synaptic plasticity, neuromodulation, dendritic computation, channel dynamics, experimental data relevant to computational issues. Reinforcement Learning and Control: exploration, planning, navigation, Q-learning, TD-learning, state estimation, dynamic programming, robotic motor control, process control, Markov decision processes. Speech and Signal Processing: speech recognition, speech coding, speech synthesis, auditory scene analysis, source separation, applications of hidden Markov models to signal processing, models of human speech perception, auditory modeling and psychoacoustics. Theory: computational learning theory, statistical physics of learning, information theory, prediction and generalization, regularization, Boltzmann machines, Helmholtz machines, decision trees, support vector machines, online learning, dynamics of learning algorithms, approximation and estimation theory, learning of dynamical systems, complexity theory. Visual Processing: image processing, image coding, object recognition, visual psychophysics, stereopsis, motion detection and tracking. For general inquiries or requests for registration material E-mail: nipsinfo at salk.edu or Fax: (619) 587-0417 Information can also be found on the NIPS web page. NIPS*99 Organizing Committee: General Chair, Sara Solla, Northwestern University; Program Chair, Todd Leen, Oregon Graduate Institute; Publications Chair, Klaus Mueller, GMD First; Tutorial Chair, Joachim Buhmann, University of Bonn; Workshops Co-Chairs, Sue Becker, McMaster University, and Rich Caruana, Carnegie Mellon University; Publicity Chair, Lee Giles, NEC Research Institute; Local Arrangements, Arun Jagota, University of California at Santa Cruz; Treasurer, Bartlett Mel, University of Southern California; Web Master, Doug Baker, Carnegie Mellon University; Government Liaison, Gary Blasdel, Harvard Medical School; Contracts, Steve Hanson, Rutgers University, and Gerry Tesauro, IBM. NIPS*99 Program Committee: Leon Bottou, AT&T Labs - Research; Gary Cottrell, University of California San Diego; Zoubin Ghahramani, University College London; Tommi Jaakkola, MIT; John Lazzaro, University of California Berkeley; Todd Leen, Oregon Graduate Institute (chair); John Moody, Oregon Graduate Institute; Barak Pearlmutter, University of New Mexico; Alexandre Pouget, University of Rochester; David Saad, Aston University; Lawrence Saul, AT&T Labs - Research; Sebastian Thrun, Carnegie Mellon University; Benjamin Van Roy, Stanford University; Paul Viola, MIT. - Please Post - __ C. Lee Giles / Computer Science / NEC Research Institute / 4 Independence Way / Princeton, NJ 08540, USA / 609-951-2642 / Fax 2488 www.neci.nec.com/~giles == From ishii at is.aist-nara.ac.jp Wed Oct 20 23:46:04 1999 From: ishii at is.aist-nara.ac.jp (Shin Ishii) Date: Thu, 21 Oct 1999 12:46:04 +0900 Subject: Paper Avialable Message-ID: <199910210346.MAA32143@axp27.aist-nara.ac.jp> Dear Connectionists, I am pleased to inform you that the following paper is available on my Web site: http://www.aist-nara.ac.jp/~ishii/publication.html We would greatly appreciate it, if you could give us any comments and suggestion. ----------------------------------------------------------------- Lambda-opt neural approaches to quadratic assignment problems Shin Ishii and Hirotaka Niitsuma to appear in Neural Computation Abstract: In this paper, we propose new analog neural approaches to combinatorial optimization problems, in particular, quadratic assignment problems (QAPs). Our proposed methods are based on an analog version of the lambda-opt heuristics, which simultaneously changes assignments for lambda elements in a permutation. Since we can take a relatively large lambda value, our new methods can achieve a middle-range search over possible solutions, and this helps the system neglect shallow local minima and escape from local minima. In experiments, we have applied our methods to relatively large-scale (N = 80-150) QAPs. Results have shown that our new methods are comparable to the present champion algorithms; for two benchmark problems, they are able to obtain better solutions than the previous champion algorithms. ----------------------------------------------------------------- Shin Ishii, Ph.D. Nara Institute of Science and Technology ATR Human Information Processing Research Laboratories From arthur.filippidis at dsto.defence.gov.au Thu Oct 21 03:36:56 1999 From: arthur.filippidis at dsto.defence.gov.au (Filippidis, Arthur) Date: Thu, 21 Oct 1999 17:06:56 +0930 Subject: Call for Papers, Special Fusion of Intelligent Systems Session at KES2000 Conference in UK, 30 Aug -1 Sept 2000 Message-ID: Special Session for KES2000 Conference. FUSION OF KNOWLEDGE-BASED INTELLIGENT SYSTEMS Architectures and Applications Organiser/Chair : Dr. Arthur Filippidis Please Circulate. Regards Dr. Arthur Filippidis System Concepts Group DSTO, Land Operations Division ph: (08) 8259 5442 fax: (08) 8259 5624 Fourth International Conference on Knowledge-Based Intelligent Engineering Systems & Allied Technologies (http://www.eng.bton.ac.uk/eee/research/kes2000/) 30 August-1 September 2000 - University of Brighton, Sussex, U.K. Special Session on : FUSION OF KNOWLEDGE-BASED INTELLIGENT SYSTEMS Architectures and Applications Call For Papers The special session aims to cover recent practical applications in the fusion of various Intelligent techniques such as, fuzzy, neural or genetic algorithms, or in combination with intelligent agents. The session also covers recent advances in knowledge-based multisensor or Information fusion architectures using a hybrid of various Intelligent techniques, in the practical areas such as for example landmine detection, target recognition, multispectral image fusion or other image fusion applications. Topics of the Special Session cover, but are not limited to: * Fusion of sensors using Intelligent techniques in areas such as target detection, landmines. * Combining the use of any Intelligent Agent technology / Fuzzy /Neural / and Genetic Algorithms to produce novel hybrid architectures used in practical applications * Multisensor Data Fusion architectures in practical applications using Intelligent techniques. * Image or Information Fusion using Intelligent techniques or Agent technology. * Multispectral fusion of Images or GIS Information using Knowledge-based techniques. INSTRUCTIONS FOR AUTHORS Interested authors must submit 2 copies of their paper to the Session Chair by post (prefered method), ( Directly to Dr. A. Filippidis, Postal Address below), or email (Word97 or earlier versions only) attachments also accepted. Submitted papers must comply to the Conference guidelines for submission (http://www.eng.bton.ac.uk/eee/research/kes2000/). All papers will be strictly refereed by at least three experts in the field. PUBLICATION The session papers will be published in the Conference Proceedings. IMPORTANT DATES Deadline for Submission: January 20th, 2000 Acceptance/Rejection Notification: March 20th, 2000 Camera-ready papers due April 10th, 2000 Session Chair: Dr. Arthur Filippidis Defence Science Technology Organisation Land Operations Division, P.O. Box 1500, Salisbury, South Australia, 5108, Australia. Email: arthur.filippidis at dsto.defence.gov.au ph: (08) 8259 5442 fax: (08) 8259 5624 From andre at icmc.sc.usp.br Fri Oct 22 04:46:39 1999 From: andre at icmc.sc.usp.br (Andre Carlos Ponce de Leon F. de Carvalho) Date: Fri, 22 Oct 1999 09:46:39 +0100 Subject: CFP for New Journal: IJCIA Message-ID: <3810246F.CBFA54AC@icmc.sc.usp.br> --------------------------------- Apologies for cross-posting --------------------------------- CALL FOR PAPERS ************************************************************* NEW JOURNAL BY IMPERIAL COLLEGE PRESS http://www.wspc.com/journals/ijcia/ijcia.html ************************************************************* INTERNATIONAL JOURNAL OF COMPUTATIONAL INTELLIGENCE AND APPLICATIONS (IJCIA) Editors-in-Chief: Andre de Carvalho, University of Guelph, Canada (until December 1999, University of Sao Paulo, Brazil) Antonio Braga, University Federal of Minas Gerais, Brazil Brijesh Verma, Griffith University, Australia Honorary Editors-in-Chief: Igor Aleksander, Imperial College, UK John Holland, University of Michigan and Santa Fe Institute, USA Lofti Zadeh, University of California at Berkeley, USA Associate Editors: Carme Torras, CSIC-UPC, Spain Harold Szu (NSWC, USA) Xin Yao, University of Birmingham, UK Witold Pedrycz, University of Alberta, Canada Editorial Board: A. Aamodt (NTNU, Norway), I. Aleksander (Imperial College, UK), J. A. Anderson (Brown University, USA), G. Cotrell (University of California San Diego, USA), Y. Dote (Muroran Institute of Technology, Japan), M. Fairhurst (University of Kent at Canterbury, UK), C. Lee Giles (NEC Research Institute, USA), M. Jamshidi (University of New Mexico, USA), J. Holland (University of Michigan and Santa Fe Institute, USA), Z. Liu (University of Melbourne, Australia), T. B. Ludermir (UFPE, Brazil), E. Mamdani (Imperial College, UK), J. Mulawka (Warsaw University of Technology, Poland), S. K. Pal (Statistical Institute of Calcutta, India), W. Pedrycz (University of Alberta, Canada), J. Shavlik (University of Wisconsin, USA), H. Szu (NSWC, USA), N. Kasabov (University of Otago, New Zealand), Y. Takefuji (Keio University , Japan), C. Torras (CSIC-UPC, Spain), G. Thimm (Nanyang Tech. Univ., Singapore), I. B. Turksen (University of Toronto, Canada), A. S. Weigend (New York University, USA), X. Yao, (University of Birmingham, UK), L. Zadeh, Univ. of Berkeley, USA Introduction: Computational Intelligence is a fastly expanding research field, attracting every year a large number of scientists, engineers and practitioners. Moreover, a growing number of companies is employing Computational Intelligence techniques to improve previous solutions and to deal with new problems. For a large number of applications, it has been realized that the combination of different techniques might lead to more efficient solutions. Aims and Scope: The International Journal of Computational Intelligence and Applications, IJCIA, is a refereed journal dedicated to the theory and applications of computational intelligence (artificial neural networks, fuzzy systems, genetic algorithms and hybrid systems). The main goal of the journal is to provide a vehicle to the scientific community and industry where ideas using two or more conventional and computational intelligence based techniques could be discussed. The journal welcomes original works in areas such as neural networks, fuzzy logic, evolutionary computation, pattern recognition, hybrid intelligent systems, symbolic machine learning, statistical models, image/audio/video compression and retrieval. The journal highly encourages new ideas, cobining two or more areas, such as neuro-fuzzy, neuro-symbolic, neuro-genetic, neuro-symbolic, neuro-pattern recognition, genetic-fuzzy, genetic-symbolic, fuzzy-symbolic, etc. Submission: Submitted papers must be in English and should not have been published elsewhere or being currently under consideration by any other journal. The original and four copies of the manuscript, not exceeding 10 pages, should be submitted to any one of the Editors-in-Chief. In addition to the hardcopies, authors should also send an electronic version of their papers in a disc (postscript or pdf formats are preferable). Authors should retain a copy of their submitted paper(s) as a guarantee for loss or damage. Addresses for submission : Andre C P L Ferreira de Carvalho Dept. of Ciencias de Computacao e Estatistica ICMC - Universidade de S?o Paulo Caixa Postal 668 CEP 13560-970 S?o Carlos, SP Brazil E-mail: andre at icmc.sc.usp.br ===> Moving in January 2000 to: <=== Dept. of Computing and Information Science University of Guelph Guelph, ON, N1G 2W1 Canada Tel.: +1 519 824-4120 Fax: +1 519 837-0323 E-mail: andre at snowhite.cis.uoguelph.ca Antonio de Padua Braga Dept. Engenharia Eletronica Campus da UFMG (Pampulha) Caixa Postal 209 30.161-970, Belo Horizonte, MG Brazil Tel.: +55 31 499 4869 (499 4848) Fax: +55 31 499 4850 E-mail: apbraga at eee.ufmg.br Brijesh Verma School of Information Technology Griffith University-Gold Coast Campus PMB 50, Gold Coast Mail Center QLD 9726 Australia Tel.: +61 755948592 Fax: +61 755948066 E-mail: b.verma at gu.edu.au "International Journal of Computational Intelligence and Applications will be distributed by World Scientific Press Company, Singapore." From abrowne at lgu.ac.uk Mon Oct 25 11:11:05 1999 From: abrowne at lgu.ac.uk (Tony Browne) Date: Mon, 25 Oct 1999 16:11:05 +0100 (GMT Daylight Time) Subject: Postdoc + PhD in UK Message-ID: Two research posts (Portsmouth/London) are available for projects in Biological Data Mining, one post-doctoral research fellow and one research studentship. A Postdoctoral Research Fellow is required for the project "Biological data mining: A comparison of neural network and symbolic techniques" funded by the joint BBSRC/EPSRC programme in bioinformatics. The Research Fellow will be based in the Centre for Molecular Design at the University of Portsmouth. This post requires either a graduate with a PhD in Computer Science, Cognitive Science, Mathematics or a related subject, with an interest in applying machine learning techniques in the field of bioinformatics; or a graduate with a PhD in the Life Sciences with a thorough understanding of bioinformatics and strong mathematical and programming skills. The successful candidate will be required to: * design, implement and test novel numerical algorithms in a high-level programming language * apply new and existing software to biological datasets * perform statistical analyses of the results The ideal candidate will possess knowledge in some of the following areas: * experience of developing neural networks in high level programming languages * strong mathematical and statistical skills (including Bayesian statistics) * an understanding of symbolic machine learning techniques * skills in Matlab, C/C++, Unix * an awareness of current developments in bioinformatics The starting salary for this post will be in the range stlg 17,606 - 19,203. A Postgraduate Research Student is required to perform research on using neural networks with biological and chemical datasets. The Research Student will be based in the School of Computing, Information Systems and Mathematics at London Guildhall University. This post requires either a graduate with a good first degree in Computer Science, Cognitive Science, Mathematics or a related subject, with an interest in applying machine learning techniques in the field of bioinformatics; or a graduate with a good first degree in the Life Sciences with a thorough understanding of bioinformatics and strong mathematical and programming skills. The successful candidate will be required to: * design, implement and test novel numerical algorithms in a high-level programming language * apply new and existing software to biological datasets * perform statistical analyses of the results A tax-free bursary is provided for this post, to a level commensurate with standard research council PHD studentships. The candidate will be encouraged to register for a PhD. As this studentship is internally funded some formal teaching will be required (typically up to 6 hours of laboratory supervision). Both of the above posts will be of 3 years duration, and will start as soon as possible. To apply please e-mail a CV (in text format) including the names (and e-mail addresses) of two referees to Dr. Antony Browne (abrowne at lgu.ac.uk) by 19th November 1999. ======================================================= Dr. Antony Browne School of Computing, Information Systems & Mathematics London Guildhall University 100 Minories London EC3 1JY, UK Tel: (+44) 0207 320 3106 Fax: (+44) 0207 320 1717 ======================================================= From bhaskar at crab.rutgers.edu Mon Oct 25 15:40:43 1999 From: bhaskar at crab.rutgers.edu (Bhaskar DasGupta) Date: Mon, 25 Oct 1999 15:40:43 -0400 (EDT) Subject: faculty position in neural computation Message-ID: <199910251940.PAA20520@bhaskar.rutgers.edu> Fellow Connectionists, Below is a faculty opening announcement in our department that would be relevant to the neural network and the learning theory community. It will also appear (sometime in future) in CACM, IEEE Computer, and the Computing Research News. Please feel free to pass it to any other interested party. My apologies if you received this multiple times or if this announcement was not relevant to you. Best regards, Bhaskar DasGupta ------------------------------------------------------------------------------ Bhaskar DasGupta Department of Computer Science Email: bhaskar at crab.rutgers.edu 319 Business & Science Building Phone: 856-225-6439 (office, in Rutgers) 3rd & Penn Street 732-445-4580 (office, in DIMACS) Rutgers University Fax: 856-225-6624 (office) Camden, NJ 08102-1656, USA URL:http://crab.rutgers.edu/~bhaskar ------------------------------------------------------------------------------ ================================================================================ Rutgers University - Camden Department of Computer Science The Department of Computer Science at Rutgers University - Camden invites applications for two tenure-track positions at the assistant professor level beginning September 2000. Candidates in all areas of computer science are encouraged to apply, particularly those with research interests in graphics and visualization, computational geometry, artificial neural networks, machine learning, networking, distributed and mobile computing, and parallel computing. A Ph.D. in Computer Science is required. Salary is contingent upon qualifications. Rutgers University stands in the top 5% of AAUP rankings of university salary and benefits. Applicants should send a curriculum vita, statement of research, statement of teaching interests/experience, and three letters of reference to: Chair, Faculty Search Committee Department of Computer Science Rutgers University Camden, NJ 08102 Review of applications will begin immediately and will continue until the positions are filled. Rutgers University is an Equal Opportunity and Affirmative Action Employer. From baolshausen at ucdavis.edu Mon Oct 25 20:54:19 1999 From: baolshausen at ucdavis.edu (Bruno Olshausen) Date: Mon, 25 Oct 1999 17:54:19 -0700 Subject: Institute for Theoretical Dynamics, UC Davis Message-ID: <3814FBBB.CB843FC8@ucdavis.edu> Dear Colleagues, The University of California, Davis, and the National Science Foundation have established a multidisciplinary training group, "Nonlinear Dynamics in Biology". The program is designed to provide graduate students with research training that will allow them to address questions in biology using computational and mathematical modelling. Specifically, our trainees focus on complex processes in cell biology, neuroscience, biofluiddynamics, ecology, and population biology that involve nonlinear dynamics. The curriculum includes training in computational, mathematical, and experimental methods. Students are mentored by faculty trainers, postdoctoral researchers, and more advanced graduate students in both formal and informal settings. Financial support is available for all graduate trainees. Admission to the graduate training program requires strong interest in biological research that involves nonlinear dynamics, acceptance to a graduate program at UC Davis, and a year-long course in calculus. We strongly encourage additional coursework in mathematics, computer science, and biology. Please alert students who might be interested in this program to the training grant web site, http://www.itd.ucdavis.edu/rtg/, for additional information and application materials. You also may contact our academic coordinator, Carole Hom (clhom at ucdavis.edu; 530-754-9117) should you or your students have questions not addressed in the information on the web site. In addition, we seek postdoctoral fellows in biofluiddynamics, cell or neurobiology, and population biology or ecology. Application materials for postdocs also can be found on the web page. Thank you for your assistance. -- Bruno A. Olshausen Phone: (530) 757-8749 Center for Neuroscience Fax: (530) 757-8827 UC Davis Email: baolshausen at ucdavis.edu 1544 Newton Ct. WWW: http://redwood.ucdavis.edu Davis, CA 95616 From luettin at idiap.ch Tue Oct 26 12:50:23 1999 From: luettin at idiap.ch (Juergen Luettin) Date: Tue, 26 Oct 1999 18:50:23 +0200 Subject: Postdocs/PhDs in Multimodal Biometrics and Multimedia Retrieval Message-ID: <199910261650.SAA21967@montfort.idiap.ch> POSTDOC and PHD POSITIONS in MULTIMODAL BIOMETRICS and MULTIMEDIA RETRIEVAL at IDIAP, Switzerland Applications are invited for several Post-Doctoral positions and PhD research studentships in the Computer Vision Group at IDIAP to work in one of the following research areas (with possible research topics): Video Indexing/Retrieval - adaptation of video content analysis methods - combination and integration of audio, video, and text information - high level video structure modelling - video abstraction/summary generation, browsing, GUI, MPEG7 Audio Indexing/Retrieval - adaptation of speech recognition systems to video (acoustic models, lexicon, language models, topic) - audio segmentation (speech, music, audio events) - speaker change detection - document indexing and retrieval, MPEG7 Text Segmentation and Recognition in Videos - detection, segmentation and recognition of text embedded in video - investigation of image properties (texture, colour, shape) - use of temporal information - integration of detection/segmentation algorithms with text recognition engine Face Analysis - face detection (in complex backgrounds, under variant pose, occlusion) - face recognition (identification/verification, robustness issues, confidence estimation, evidence integration) Speaker Verification - client model adaptation - decision strategy - robustness to microphone/channel characteristics - discriminant analysis Classifier Combination - classifier confidence estimation - classifier combination techniques - evidence integration - application to multimodal biometrics, video retrieval The successful candidates will participate in the following important and highly competitive European and national projects: o BANCA: Biometric Access Control for Networked and e-Commerce EU project, lead by Matra Nortel Communication http://www.idiap.ch/vision/banca.html o ASSAVID: Automatic Segmentation and Semantic Annotation of Sports Videos EU project, lead by Sony UK http://www.idiap.ch/vision/assavid.html o VOCR: Text Segmentation and Recognition for Video Retrieval Swiss National Science Foundation project http://www.idiap.ch/vision/vocr.html Several tasks will be performed in close collaboration with other project partners including universities, industries, and end-users. REQUIREMENTS Postdoctoral candidates should possess a relevant PhD degree and doctoral students an M.S. degree (or equivalent) in Computer Science or a related area and should have a strong background in mathematics, signal processing, and pattern recognition. ABOUT IDIAP IDIAP is a semi-private research institute, that conducts basic and applied research in artificial intelligence. The main research activities are in machine learning, speech processing, and computer vision. Research in the rapidly growing computer vision group focuses on topics in multimedia processing including multimodal biometrics, multimodal speech recognition, and document analysis and recognition (See http://www.idiap/vision). Doctoral students at IDIAP are usually registered at a Swiss university or Federal Institute of Technology, often EPFL, that awards the PhD degree. LOCATION IDIAP is located in the town of Martigny in Valais, a scenic region in the south of Switzerland, surrounded by some of the highest mountains in Europe which offer some of the best skiing, hiking, and climbing. It is within close proximity to Lausanne, and lake Geneva, and centrally located for travel to other parts of Europe. HOW TO APPLY Prospective candidates should send a detailed CV, including 2 names and addresses of referees, research interest, and earliest possible starting date, preferably by email to Dr. Juergen Luettin Head, Computer Vision Group IDIAP, CP 592, 1920 Martigny, Switzerland Email: Luettin at idiap.ch Tel: +41 27 721 77 27 Fax: +41 27 721 77 12 From ken at phy.ucsf.EDU Wed Oct 27 05:56:53 1999 From: ken at phy.ucsf.EDU (Ken Miller) Date: Wed, 27 Oct 1999 02:56:53 -0700 (PDT) Subject: Paper available: Review of Orientation Selectivity Development and Models Message-ID: <14358.52325.953582.526768@coltrane.ucsf.edu> The following paper is now available at ftp://ftp.keck.ucsf.edu/pub/ken/or-review.ps.gz (compressed postscript) or http://www.keck.ucsf.edu/~ken (click on 'Publications') This is a preprint of an article that appeared as Journal of Neurobiology 41:44-57 (1999): http://www3.interscience.wiley.com/cgi-bin/abstract?ID=66000451 ------------------------------ Is the Development of Orientation Selectivity Instructed by Activity? Kenneth D. Miller, Ed Erwin and Andrew Kayser Dept. of Physiology, UCSF ABSTRACT: Is the development of orientation selectivity in visual cortex instructed by the patterns of neural activity of input neurons? We review evidence as to the role of activity, review models of activity-instructed development, and discuss how these models can be tested. The models can explain the normal development of simple cells with binocularly matched orientation preferences, the effects of monocular deprivation and reverse suture on the orientation map, and the development of a full intracortical circuit sufficient to explain mature response properties including the contrast-invariance of orientation tuning. Existing experiments are consistent with the models, in that (1) selective blockade of ON-center ganglion cells, which will degrade or eliminate the information predicted to drive development of orientation selectivity, in fact prevents development of orientation selectivity; and (2) the spontaneous activities of inputs serving the two eyes are correlated in the lateral geniculate nucleus at appropriate developmental times, as was predicted to be required to achieve binocular matching of preferred orientations. However, definitive tests remain to be done to (1) firmly establish the instructive rather than simply permissive role of activity and (2) determine whether the retinotopically- and center-type-specific patterns of activity predicted by the models actually exist. We conclude by critically examining alternative scenarios for the development of orientation selectivity and maps, including the idea that maps are genetically pre-specified. Ken Kenneth D. Miller telephone: (415) 476-8217 Dept. of Physiology fax: (415) 476-4929 UCSF internet: ken at phy.ucsf.edu 513 Parnassus www: http://www.keck.ucsf.edu/~ken San Francisco, CA 94143-0444 From mw at stat.Duke.EDU Wed Oct 27 09:27:48 1999 From: mw at stat.Duke.EDU (Mike West) Date: Wed, 27 Oct 1999 09:27:48 -0400 Subject: No subject Message-ID: <19991027092748.E26398@isds.duke.edu> ------------------------------------------------------------------------- ------------------------------------------------------------------------- MITCHELL PRIZE 2000: ANNOUNCEMENT AND SOLICITATION http://www.stat.duke.edu/sites/mitchell.html The Mitchell Prize is awarded in recognition of an outstanding paper that describes how a Bayesian analysis has solved an important applied problem. The 2000 Prize includes an award of $1000 and a commemorative plaque, and will be announced and presented at the ISBA 2000 meeting in Crete (May 28-June 1 2000). The Mitchell Prize is named for Toby J. Mitchell and was established by his friends and colleagues following his death from leukemia in 1993. Toby was a Senior Research Staff Member at Oak Ridge National Laboratory throughout his career, will leaves of absence spent at the University of Wisconsin and at the National Institute of Environmental Health Sciences. Toby won the Snedecor Award in 1978 (with co-author Bruce Turnbull), made incisive contributions to statistics, especially in biometry and engineering applications, and was a marvelous collaborator and an especially thoughtful scientist. Toby was a dedicated Bayesian, hence the focus of the prize. This is the fourth Mitchell Prize, the first three having been awarded in 1994, 1997 and 1999. Since 1999 the Prize is awarded annually under the cosponsorship of the ASA Section on Bayesian Statistical Science (SBSS), the International Society for Bayesian Analysis (ISBA), and the Mitchell Prize Founders' Committee. The awarding of the Mitchell Prize is governed by the Mitchell Prize charter, established in 1999 (and available at the Mitchell Prize web site, noted above). Under this charter, the sponsors annually establish a selection committee; the 2000 Prize selection committee members are Gary Koop, Henry Wynn and Mike West (chair). To be eligible for the 2000 Prize, a paper will either have appeared in a refereed journal or refereed conference proceedings since January 1 1998, or be scheduled for future publication in a refereed outlet. Candidate papers will be accepted from nominators and from authors. In reviewing submissions, emphasis will be placed on evidence that the application has truly benefited from a Bayesian analysis respecting the individual character of the problem at hand. There is no restriction as to approach taken, except that it be Bayesian in some sense, and that it carefully and appropriately justifies models, priors and methodologies adopted. To be considered for this year's Prize, please submit the following: * FOUR reprints or copies of the manuscript * A cover letter, with two copies, containing the following elements: -- A brief statement of the impact of the work -- Contact information for the authors and nominator (if not an author) -- Full email and postal addresses, plus telephone numbers, for TWO individuals who can be contacted for an evaluation of the importance of the work in the applied field. The named individuals should be experts in the applied field in question, but must not be either statisticians or coauthors/collaborators of those named on the submission. Submissions should be mailed to Mike West, Mitchell Prize Selection Committee Chair Institute of Statistics & Decision Sciences Duke University Durham, NC 27708-0251 USA Entries must be received at this address by JANUARY 31st 2000 in order to receive consideration. Visit the web site listed above to learn more about the Mitchell Prize and the sponsoring organizations. ------------------------------------------------------------------------ ------------------------------------------------------------------------ From Frederic.Alexandre at loria.fr Thu Oct 28 10:30:37 1999 From: Frederic.Alexandre at loria.fr (Frederic Alexandre) Date: Thu, 28 Oct 1999 16:30:37 +0200 Subject: PostDoc Position at LORIA-INRIA France Message-ID: <38185E0D.6257@loria.fr> LORIA/INRIA Lorraine computer science laboratory in Nancy, France One postdoctoral position is available for developping image processing systems in the framework of computational neurosciences applied to robotics, in Nancy, France, from January to December 2000. Our team: the CORTEX team is developping connectionist models, inspired from biology, for perception, reasoning and autonomous behavior. Belonging to a computer science lab, our main goal is to propose effective models for robotics, speech and image processing. The position will consist in helping PhD students to build the visual processing system that have to be interfaced to their biologically inspired models. This system will be itself biologically inspired or will simply include classical image processing algorithms. The work thus implies developping an interface, for a Koala autonomous robot, between a video camera and our cortically inspired models. The computer scientist candidate: he/she should be well-trained in image processing and connectionist modeling. Experience in robotics, autonomous behavior and neurosciences would be highly appreciated. For further information, contact: ---------------------------------------------------------------------------- Frederic ALEXANDRE Tel: (+33/0) 3 83 59 20 53 INRIA-Lorraine/LORIA-CNRS Fax: (+33/0) 3 83 41 30 79 BP 239 E-mail: falex at loria.fr 54506 Vandoeuvre-les-Nancy Cedex http://www.loria.fr/~falex FRANCE ---------------------------------------------------------------------------- From maass at igi.tu-graz.ac.at Fri Oct 29 17:04:49 1999 From: maass at igi.tu-graz.ac.at (Wolfgang Maass) Date: Fri, 29 Oct 1999 23:04:49 +0200 Subject: Neural nets with dynamic synapses and nonlinear filters Message-ID: <381A0BF1.B4117AF6@igi.tu-graz.ac.at> The following paper is now online available: Neural Systems as Nonlinear Filters Wolfgang Maass and Eduardo D.Sontag Technische Univ. Graz Rutgers University Abstract: Experimental data show that biological synapses behave quite differently from the symbolic synapses in all common artificial neural network models. Biological synapses are dynamic, i.e., their ``weight'' changes on a short time scale by several hundred percent in dependence of the past input to the synapse. In this article we address the question how this inherent synaptic dynamics-- which should not be confused with long term ``learning'' -- affects the computational power of a neural network. In particular we analyze computations on temporal and spatio-temporal patterns, and we give a complete mathematical characterization of all filters that can be approximated by feedforward neural networks with dynamic synapses. It turns out that even with just a single hidden layer such networks can approximate a very rich class of nonlinear filters: all filters that can be characterized by Volterra series. This result is robust with regard to various changes in the model for synaptic dynamics. Our characterization result provides for all nonlinear filters that are approximable by Volterra series a new complexity hierarchy which is related to the cost of implementing such filters in neural systems. The article will appear in Neural Computation. It is online available as # 107 from http://www.tu-graz.ac.at/igi/maass/#Publications and from http://www.math.rutgers.edu/~sontag/FTP_DIR/spiking.ps.gz From reggia at cs.umd.edu Fri Oct 29 12:48:09 1999 From: reggia at cs.umd.edu (James A. Reggia) Date: Fri, 29 Oct 1999 12:48:09 -0400 (EDT) Subject: postdoc position, computational neuroscience and language Message-ID: <199910291648.MAA28723@avion.cs.umd.edu> Post-doctoral Fellowships in the Cognitive Neuroscience of Language and its Disorders Two-year National Research Service Award fellowships are available at the University of Maryland School of Medicine, in Baltimore, Maryland. Training opportunities will provide experience in the application of contemporary research methods (including computational modeling, cognitive neuropsychology, event-related potentials and functional neuroimaging) to the topic of normal and disordered language processing. Applicants with doctoral degrees in related basic science areas (computer sciene, neuroscience, linguistics, cognitive psychology, etc.) and clinical disciplines (speech/language pathology; clinical neuropsychology) are invited to apply. Applicants must be U.S. citizens or permanent residents to be considered, under the terms of the NRSA program. Inquiries may be directed to Rita Berndt at rberndt at umaryland.edu or to Jim Reggia at reggia at cs.umd.edu . To apply, send HARD COPIES of your C.V., the names and addresses of three referees, and a statement of research interests and career goals to: Rita S. Berndt, Ph.D. Department of Neurology University of Maryland School of Medicine 22 South Greene Street Baltimore, Maryland 21201 USA From Nigel.Goddard at ed.ac.uk Mon Oct 25 22:38:42 1999 From: Nigel.Goddard at ed.ac.uk (Nigel Goddard) Date: Tue, 26 Oct 1999 03:38:42 +0100 Subject: Jobs in Neural Simulation Message-ID: <38151432.1BEE906A@ed.ac.uk> POSTDOCTORAL RESEARCH FELLOW & RESEARCH PROGRAMMER IN NEURAL SIMULATION METHODS The NEOSIM project, funded by NIMH and NSF, is developing simulation environments and associated software tools for modelling of brain processes. The project has computer science and neuroscience research goals, and aims to develop distributable, portable, parallel software for the computational neuroscience community. We seek a postdoctoral Research Fellow and a Research Programmer to augment the existing multidisciplinary, multinational team. Both posts may require some travel to other EU countries or the US. The Research Fellow will conduct original research related to brain modeling, brain data analysis or parallel simulation techniques and will contribute design, documentation, programming or other relevant expertise to the software development goals of the project. The Research Programmer will be involved in design, documentation, implementation, testing, dissemination, maintenance and development of the NEOSIM framework and will also be responsible for the installation and maintenance of software packages on the research computers, and may be involved in the specification, installation and system administration of research computers including small parallel platforms. Tenable for up to 3 years subject to annual renewal and further renewal beyond 3 years subject to success in attracting further funding. Salary scale: ?16,286 - ?24,479 p.a. Further details are available at http://anc.ed.ac.uk/neosim/rafurthpart.html Informal enquiries to: Dr. Nigel Goddard (Nigel.Goddard at ed.ac.uk) Sites: http://anc.ed.ac.uk/neosim, http://www.informatics.ed.ac.uk, http://www.personnel.ed.ac.uk/recruit.htm Further particulars, including details of the application procedure should be obtained from the Personnel Department, 1 Roxburgh Street, Edinburgh EH8 9TB, tel: +44 131-650-2511 (24 hour answering service); or see the web site above Closing Date: 24th November 1999. Interviews expected 9th-13th December. -- ======================= Dr. Nigel Goddard Institute for Adaptive and Neural Computation Division of Informatics University of Edinburgh 5 Forrest Hill Edinburgh EH1 2QL Scotland Telephone: +44 131 650 3087 email: Nigel.Goddard at ed.ac.uk web: http://anc.ed.ac.uk/~ngoddard Fax (paper): +44 131 650 6899 eFAX (email): +1 603 698 5854 ======================= From Volker.Tresp at mchp.siemens.de Fri Oct 1 09:55:33 1999 From: Volker.Tresp at mchp.siemens.de (Volker Tresp) Date: Fri, 01 Oct 1999 15:55:33 +0200 Subject: Two papers on probabilistic approximations to Bayesian networks Message-ID: <37F4BD55.E7D5DDCF@mchp.siemens.de> We would like to announce the availability of two papers. In the first paper we apply different mixture approximations to Bayesian networks with the goal of obtaining insight into the modeled domain. One of the approximations is obtained via a mixture of mean field solutions. The latter approach and a general formulation of the mean field approximation applicable to an arbitrary Bayesian network can be found in the second paper. ------------------------------------------------------------------------- Mixture Approximations to Bayesian Networks Volker Tresp, Michael Haft and Reimar Hofmann Siemens AG, Corporate Technology Neural Computation Dept. Information and Communications Otto-Hahn-Ring 6, 81730 Munich, Germany Published in Laskey, K. B., Prade, H., (Hrsg.), Uncertainty in Artificial Intelligence, Proceedings of the Fifteenth Conference, Morgan Kaufmann Publishers, 1999, pp. 639-646 Structure and parameters in a Bayesian network uniquely specify the probability distribution of the modeled domain. The locality of both structure and probabilistic information are the great benefits of Bayesian networks and require the modeler to only specify local information. On the other hand this locality of information might prevent the modeler ---and even more any other person--- from obtaining a general overview of the important relationships within the domain. The goal of the work presented in this paper is to provide an ``alternative'' view on the knowledge encoded in a Bayesian network which might sometimes be very helpful for providing insights into the underlying domain. The basic idea is to calculate a mixture approximation to the probability distribution represented by the Bayesian network. The mixture component densities can be thought of as representing typical scenarios implied by the Bayesian model, providing intuition about the basic relationships. As an additional benefit, performing inference in the approximate model is very simple and intuitive and can provide additional insights. The computational complexity for the calculation of the mixture approximations critically depends on the measure which defines the distance between the probability distribution represented by the Bayesian network and the approximate distribution. Both the KL-divergence and the backward KL-divergence lead to inefficient algorithms. Incidentally, the latter is used in recent work on mixtures of mean field solutions to which the work presented here is closely related. We show, however, that using a mean squared error cost function leads to update equations which can be solved using the junction tree algorithm. We conclude that the mean squared error cost function can be used for Bayesian networks in which inference based on the junction tree is tractable. For large networks, however, one may have to rely on mean field approximations. http://www7.informatik.tu-muenchen.de/~hofmannr/uai99_abstr.html ----------------------------------------------------------------------- ----------------------------------------------------------------------- Model-Independent Mean Field Theory as a Local Method for Approximate Propagation of Information M.~Haft, R.~Hofmann and V.~Tresp Corporate Technology, Department: Information and Communications Siemens AG, 81730 M\"unchen, Germany Published in Network: Computation in Neural Systems, 10, 1999, pp. 93-105 (based on a TR of 1997). We present a systematic approach to mean field theory (MFT) in a general probabilistic setting without assuming a particular model. The mean field equations derived here may serve as a local and thus very simple method for approximate inference in probabilistic models such as Boltzmann machines or Bayesian networks. Our approach is `model-independent' in the sense that we do not assume a particular type of dependencies; in a Bayesian network, for example, we allow arbitrary tables to specify conditional dependencies. In general, there are multiple solutions to the mean field equations. We show that improved estimates can be obtained by forming a weighted mixture of the multiple mean field solutions. Simple approximate expressions for the mixture weights are given. The general formalism derived so far is evaluated for the special case of Bayesian networks. The benefits of taking into account multiple solutions are demonstrated by using MFT for inference in a small and in a very large Bayesian network. The results are compared to the exact results. http://www7.informatik.tu-muenchen.de/~hofmannr/network99_abstr.html ----------------------------------------------------------------------- From terry at salk.edu Sun Oct 3 23:48:17 1999 From: terry at salk.edu (Terry Sejnowski) Date: Sun, 3 Oct 1999 20:48:17 -0700 (PDT) Subject: NEURAL COMPUTATION 11:8 Message-ID: <199910040348.UAA07616@dax.salk.edu> *** Volume 12 of Neural Computation will comprise 12 monthly issue *** Neural Computation - Contents - Volume 11, Number 8 - November 15, 1999 ARTICLES Detecting And Estimating Signals In Noisy Cable Structure: I. Neuronal Noise Sources Amit Manwani and Christof Koch Detecting And Estimating Signals In Noisy Cable Structures: II. Information-Theoretic Analysis Amit Manwani and Christof Koch Population Dynamics of Spiking Neurons: Fast Transients, Asynchronous States, and Locking Wulfram Gerstner Correctness of Local Probability in Graphical Models with Loops Yair Weiss NOTES Natural Gradient Learning For Over-and-Under-Complete Bases In ICA Shun-ichi Amari Combined 5x2cv F Test For Comparing Supervised Classification Learning Algorithms Ethem Alpaydin LETTER Adaptive Neural Coding Dependent on the Time-varying of the Somatic Input Current Jonghan Shin, Christof Koch, and Rodney Douglas A Reinforcement Learning Approach To On-Line Clustering Aristidis Likas Replicator Equations, Maximal Cliques, And Graph Isomorphism Marcello Pelillo Independent Component Analysis: a Flexible Non-Linearity and Decorrelating Manifold Approach Richard Everson and Stephen Roberts On The Design Of BSB Neural Associative Memories Using Semidefinite Programming Jooyoung Park, Hyuk Cho and Daihee Park How To Design A Connectionist Holistic Parser Ho K. S. Edward and Chan L. Wan A Unified Analysis Of Value-Function-Based Reinforcement- Learning Algorithms Csaba Szepesvari, and Michael Littman Neuronal Regulation: A Mechanism For Synaptic Pruning During Brain Maturation Gal Chechik, Isaac Meilijson and Eytan Ruppin Comparison of SOM Point Densities Based on Different Criteria Teuvo Kohonen ----- NOTE: Neural Computation is now on-line and issues starting with 11:1 are available to all free for a trial period: ON-LINE - http://neco.mitpress.org/ http://intl-neco.mitpress.org (Europe and Asia) ABSTRACT http://intl-neco.mitpress.orgECO/ SUBSCRIPTIONS - 1999 - VOLUME 11 - 8 ISSUES USA Canada* Other Countries Student/Retired $50 $53.50 $84 Individual $82 $87.74 $116 Institution $302 $323.14 $336 * includes 7% GST (Back issues from Volumes 1-10 are regularly available for $28 each to institutions and $14 each for individuals. (Back issues from Volumes 1-10 are regularly available for $28 each to institutions and $14 each for individuals. Add $5 for postage per issue outside USA and Canada. Add +7% GST for Canada.) MIT Press Journals, 5 Cambridge Center, Cambridge, MA 02142-9902. Tel: (617) 253-2889 FAX: (617) 258-6779 mitpress-orders at mit.edu ----- From dimi at ci.tuwien.ac.at Mon Oct 4 10:06:00 1999 From: dimi at ci.tuwien.ac.at (Evgenia Dimitriadou) Date: Mon, 4 Oct 1999 16:06:00 +0200 (CEST) Subject: CI BibTeX Collection -- Update Message-ID: The following volumes have been added to the collection of BibTeX files maintained by the Vienna Center for Computational Intelligence: IEEE Transactions on Evolutionary Computation, Volumes 2/3-3/3 IEEE Transactions on Neural Networks, Volumes 9/3-10/5 Machine Learning, Volumes 31-37 Neural Computation, Volumes 11/4-11/6 Neural Networks, Volumes 12/3-12/8 Neural Processing Letters, Volumes 9/3-10/1 Most files have been converted automatically from various source formats, please report any bugs you find. The complete collection can be downloaded from http://www.ci.tuwien.ac.at/docs/ci/bibtex_collection.html ftp://ftp.ci.tuwien.ac.at/pub/texmf/bibtex/ Best, Vivi ************************************************************************ * Evgenia Dimitriadou * ************************************************************************ * Institut für Statistik * Tel: (+43 1) 58801 10773 * * Technische Universität Wien * Fax: (+43 1) 58801 10798 * * Wiedner Hauptstr. 8-10/1071 * Evgenia.Dimitriadou at ci.tuwien.ac.at * * A-1040 Wien, Austria * http://www.ci.tuwien.ac.at/~dimi* ************************************************************************ From evansdj at aston.ac.uk Mon Oct 4 07:27:35 1999 From: evansdj at aston.ac.uk (DJ EVANS) Date: Mon, 04 Oct 1999 12:27:35 +0100 Subject: RESEARCH FELLOWSHIP: Pollution-in-Water: Detection, tracking and classification Message-ID: <37F88F27.871A8700@aston.ac.uk> Dear All, Please reply to Prof D. Lowe (d.lowe at aston.ac.uk). Thanks, David Evans. ******************************************************************* Neural Computing Research Group ------------------------------- School of Engineering and Applied Sciences Aston University, Birmingham, UK RESEARCH FELLOWSHIP ------------------- Pollution-in-Water: Detection, tracking and classification ---------------------------------------------------------- The Neural Computing Research Group at Aston is looking for a highly motivated individual for a 30 month postdoctoral research position in the area of automatic video analysis to detect, classify and track near-shore sea-borne pollutants. This project will be part of a Framework V European Union collaborative project on an environmental early warning monitoring initiative. Potential candidates should have strong statistical and computational skills, preferably with a background related to neural networks and image analysis. Further information on the project can be obtained from http://www.ncrg.aston.ac.uk/ Conditions of Service --------------------- Salaries will be up to point 9 on the RA 1A scale. The salary scale is subject to review and annual increments. How to Apply ------------ If you wish to be considered for this Fellowship, please send a full CV and publications list, including full details and grades of academic qualifications, together with the names of 3 referees, to: Prof David Lowe Neural Computing Research Group School of Engineering and Applied Sciences Aston University Birmingham B4 7ET, U.K. Tel: 0121 333 4631 Fax: 0121 333 4586 e-mail: d.lowe at aston.ac.uk e-mail submission of postscript/latex files is welcome. Closing date: 13 December, 1999. From ken at phy.ucsf.EDU Tue Oct 5 06:46:20 1999 From: ken at phy.ucsf.EDU (Ken Miller) Date: Tue, 5 Oct 1999 03:46:20 -0700 (PDT) Subject: UCSF Postdoctoral/Graduate Fellowships in Theoretical Neurobiology Message-ID: <14329.55036.61411.427912@coltrane.ucsf.edu> FULL INFO: http://www.sloan.ucsf.edu/sloan/sloan-info.html PLEASE DO NOT USE 'REPLY'; FOR MORE INFO USE ABOVE WEB SITE OR CONTACT ADDRESSES GIVEN BELOW The Sloan Center for Theoretical Neurobiology at UCSF solicits applications for pre- and post-doctoral fellowships, with the goal of bringing theoretical approaches to bear on neuroscience. Applicants should have a strong background and education in mathematics, theoretical or experimental physics, or computer science, and commitment to a future research career in neuroscience. Prior biological or neuroscience training is not required. The Sloan Center offers opportunities to combine theoretical and experimental approaches to understanding the operation of the intact brain. Young scientists with strong theoretical backgrounds will receive scientific training in experimental approaches to understanding the operation of the intact brain. They will learn to integrate their theoretical abilities with these experimental approaches to form a mature research program in integrative neuroscience. The research undertaken by the trainees may be theoretical, experimental, or a combination. Resident Faculty of the Sloan Center and their research interests include: Allison Doupe: Development of song recognition and production in songbirds Stephen Lisberger: Learning and memory in a simple motor reflex, the vestibulo-ocular reflex, and visual guidance of smooth pursuit eye movements by the cerebral cortex Michael Merzenich: Experience-dependent plasticity underlying learning in the adult cerebral cortex, and the neurological bases of learning disabilities in children Kenneth Miller: Circuitry of the cerebral cortex: its structure, self-organization, and computational function (primarily using cat primary visual cortex as a model system) Roger Nicoll: Synaptic and cellular mechanisms of learning and memory in the hippocampus Christoph Schreiner: Cortical mechanisms of perception of complex sounds such as speech in adults, and plasticity of speech recognition in children and adults Michael Stryker: Mechanisms that guide development of the visual cortex There are also a number of visiting faculty, including Larry Abbott, Brandeis University; Bill Bialek, NEC Research; Sebastian Seung, MIT; David Sparks, Baylor University; Steve Zucker, Yale University. TO APPLY, please send a curriculum vitae, a statement of previous research and research goals, up to three relevant publications, and have two letters of recommendation sent to us. The application deadline is February 1, 2000. Send applications to: Steve Lisberger Sloan Center for Theoretical Neurobiology at UCSF Department of Physiology University of California 513 Parnassus Ave. San Francisco, CA 94143-0444 PRE-DOCTORAL applicants with strong theoretical training may seek admission into the UCSF Neuroscience Graduate Program as a first-year student. Applicants seeking such admission must apply by Jan. 5, 2000 to be considered for fall, 2000 admission. Application materials for the UCSF Neuroscience Program may be obtained from http://www.neuroscience.ucsf.edu/neuroscience/admission.html or from Cindy Kelly Neuroscience Graduate Program Department of Physiology University of California San Francisco San Francisco, CA 94143-0444 neuroscience at phy.ucsf.edu Be sure to include your surface-mail address. The procedure is: make a normal application to the UCSF Neuroscience program; but also alert the Sloan Center of your application, by writing to Steve Lisberger at the address given above. If you need more information: -- Consult the Sloan Center WWW Home Page: http://www.sloan.ucsf.edu/sloan -- Send e-mail to sloan-info at phy.ucsf.edu -- See also the home page for the W.M. Keck Foundation Center for Integrative Neuroscience, in which the Sloan Center is housed: http://www.keck.ucsf.edu/ From danr at cs.uiuc.edu Wed Oct 6 10:01:20 1999 From: danr at cs.uiuc.edu (Dan Roth) Date: Wed, 06 Oct 1999 09:01:20 -0500 Subject: Software release: SNoW Learning Architecture References: <199910051432.PAA00954@tahiti> Message-ID: <37FB5630.5AEC22AA@cs.uiuc.edu> ---------------------------------------------------------------------- Software release: SNoW Learning Architecture 2.0 Cognitive Computation Group University of Illinois at Urbana/Champaign http://L2R.cs.uiuc.edu/~cogcomp/ ---------------------------------------------------------------------- The Cognitive Computation Group at the University of Illinois at Urbana/Champaign announces the release of the SNoW Learning Architecture (version 2.0). SNoW is a learning program that can be used as a general purpose multi-class classifier and is specifically tailored for learning in the presence of a very large number of features. The learning architecture is a sparse network of linear units over a pre-defined or incrementally acquired feature space. Several update rules may be used - sparse variations of the Winnow update rule, Perceptron, or naive Bayes. SNoW has been used successfully in several applications in the natural language and visual processing domains; the release is meant to be used only for research purposes, with the hope that it can be a useful research tool for studying learning in these domains. Feedback of any sort is welcome. You are invited to download the SNoW package for educational or non-commercial research purposes. When downloading the package you are asked to register and express your agreement with the license terms, under the university of Illinois guidelines. SNoW is not shareware or public domain software. The SNoW software package can be downloaded by following the `Software' link under the Cognitive Computation group home page at http://L2R.cs.uiuc.edu/~cogcomp The SNoW package contains the following: - Source code (C++) with a Makefile. - A user guide (UIUC Tech Report, UIUC-DCS-R-99-210) containing brief description of the architecture and algorithms, detailed descriptions of the command-line options, and a brief tutorial. - The text of the license agreement. Two packages are available, one for UNIX system (that should be easily installed also on most Linux systems) and a second for NT systems. The User guide (UIUC Tech report UIUC-DCS-R-99-210, by Andrew J. Carlson, Chad M. Cumby, Jeff L. Rosen and Dan Roth) can be downloaded directly from: http://L2R.cs.uiuc.edu/~danr/Papers/userguide.ps.gz Papers related to SNoW can be found at http://L2R.cs.uiuc.edu/~danr/snow.html For comments and bug reports relating to SNoW, please send mail to SNoW at cs.uiuc.edu ---------------------------------------------------------------------- Dan Roth Department of Computer Science, University of Illinois, Urbana/Champaign 1304 W. Springfield Ave. Urbana IL 61801 Phone: (217) 244-7068 (217) 244-6813 (Sec) Fax: +(217) 244-6500 e-mail: danr at cs.uiuc.edu http://L2R.cs.uiuc.edu/~danr ---------------------------------------------------------------------- From steve at cns.bu.edu Wed Oct 6 20:25:03 1999 From: steve at cns.bu.edu (Stephen Grossberg) Date: Wed, 6 Oct 1999 20:25:03 -0400 Subject: How The Basal Ganglia Learn to Selectively Respond to Unexpected Rewarding Cues Message-ID: The following article can be read at http://cns-web/bu.edu/Profiles/Grossberg HOW THE BASAL GANGLIA USE PARALLEL EXCITATORY AND INHIBITORY LEARNING PATHWAYS TO SELECTIVELY RESPOND TO UNEXPECTED REWARDING CUES. Brown, J., Bullock, D., and Grossberg, S. (1999). Journal of Neuroscience, in press. After classically conditioned learning, dopaminergic cells in the substantia nigra pars compacta (SNc) respond immediately to unexpected conditioned stimuli (CS) but omit formerly seen responses to expected unconditioned stimuli, notably rewards. These cells play an important role in reinforcement learning. A neural model explains the key neurophysiological properties of these cells before, during, and after conditioning, as well as related anatomical and neurophysiological data about the pedunculo-pontine tegmental nucleus (PPTN), lateral hypothalamus, ventral striatum, and striosomes. The model proposes how two parallel learning pathways from limbic cortex to the SNc, one devoted to excitatory conditioning (through the ventral striatum, ventral pallidum, and PPTN) and the other to adaptively timed inhibitory conditioning (through the striosomes), control SNc responses. The excitatory pathway generates CS-induced excitatory SNc dopamine bursts. The inhibitory pathway prevents dopamine bursts in response to predictable reward-related signals. When expected rewards are not received, striosomal inhibition of SNc that is unopposed by excitation results in a phasic drop in dopamine cell activity. The adaptively timed inhibitory learning uses an intracellular spectrum of timed responses that is proposed to be similar to adaptively timed cellular mechanisms in the hippocampus and the cerebellum. These mechanisms are proposed to include metabotropic glutamate receptor-mediated Ca2+ spikes that occur with different delays in striosomal cells. A dopaminergic burst in concert with a Ca2+ spike is proposed to potentiate inhibitory learning. The model provides a biologically predictive alternative to temporal difference (TD) conditioning models and explains substantially more data than alternative models. Keywords: dopamine, substantia nigra, reward, basal ganglia, conditioning, pedunculopontine tegmental nucleus, lateral hypothalamus, striosomes, adaptive timing Preliminary version appears as Boston University Technical Report, CAS/CNS-TR-99-011. Download this paper as: Gzipped Postscript BroBulGro99.ps.gz (152Kb) or PDF (BroBulGro99.pdf) From gary at cs.ucsd.edu Wed Oct 6 21:42:52 1999 From: gary at cs.ucsd.edu (Gary Cottrell) Date: Wed, 6 Oct 1999 18:42:52 -0700 (PDT) Subject: Jobs at UCSD Message-ID: <199910070142.SAA20655@gremlin.ucsd.edu> Hi folks - This is a "heads up" that the CSE Department at UCSD has several faculty positions at Junior AND Senior levels that we are going to try to fill this coming year. While the ad specifies many areas, among the ones listed are machine learning, data mining, bioinformatics, and computational biology. The goal is to hire the best candidates, regardless of area. If you are one of those, please apply! See the web page: http://www-cse.ucsd.edu/cse/Academic.Jobs.html The deadline is January 15, 2000. The CSE Department at UCSD is consistently ranked among the top 20 in the US, and has made several key hires in recent years in theory, networking, distributed systems. UCSD has been ranked the number 1 public science university in the US using only objective measures. There is a broad group of people around the university interested in various standard combinations of the words Neural, Computation, Cognition, and Science. Plus, we have the best weather - the beach party last year was on Halloween! cheers, gary Gary Cottrell 858-534-6640 FAX: 858-534-7029 Faculty Assistant Chet Frost: 858-822-3286 Computer Science and Engineering 0114 IF USING FED EX INCLUDE THE FOLLOWING LINE: "Only connect" 3101 Applied Physics and Math Building University of California San Diego -E.M. Forster La Jolla, Ca. 92093-0114 Email: gary at cs.ucsd.edu or gcottrell at ucsd.edu Home page: http://www-cse.ucsd.edu/~gary/ From D.Willshaw at cns.ed.ac.uk Thu Oct 7 06:46:55 1999 From: D.Willshaw at cns.ed.ac.uk (David Willshaw) Date: Thu, 7 Oct 1999 11:46:55 +0100 (BST) Subject: NETWORK: Computation in Neural Systems Message-ID: <14332.31263.408603.380963@gargle.gargle.HOWL> `NETWORK: Computation in Neural Systems' has changed its scope. A shortened version of the Editorial which will appear in the next issue follows. CHANGES AT `NETWORK: COMPUTATION IN NEURAL SYSTEMS' Journals such as `NETWORK: Computation in Neural Systems' have exploited the essential ambiguity in such terms as `computation in neural systems' by appealing both to those interested in the nervous system and those interested in neurally-inspired artefacts and algorithms. However, with the rapid expansion seen in both areas of research, there would come a time when it would be practicable for the journal to focus onto one of these research areas. This time has come. Given the underlying interest within `NETWORK' in the nervous system, it is natural that the journal now focusses on `computational neuroscience'. The journal provides a forum for integrating theoretical and experimental findings in computational neuroscience across relevant interdisciplinary boundaries. It aims to make theor- etical results and methods accessible to neurobiologists, psychologists and cognitive scientists. Similarly, the rapidly accumulating empirical data in the neurobiological, psych- ological and cognitive domains will enable theorists to stimulate a synthesis or provoke new models. This is the new scope of `NETWORK: Computation in Neural Systems' which is published in every issue. `NETWORK' spans the disciplines of mathematics, physics, computer science, psychology, cognitive science, medicine, neurobiology, amongst others. Work on `computation in neural systems' as defined here refers to work on theoretical and computational aspects of the development and functioning of the nervous system, which can be at the level of networks of neurons or at the cellular or the subcellular level; or reporting new experimental neuroscience findings which are presented within the context of a theory or model; or reporting the application of more abstract, neural network models to problems at the psychological, cognitive and linguistic levels; or developing and analysing novel theories and models which have a strong and direct inspiration from natural systems such as neurobiological systems. In practical terms, the only change discernable in the contents of the journal will be that work on neural networks without a strong biological application or inspiration other than that expressed in the term `neural networks' now falls outside the scope of the journal. Recent Organisational Changes at `NETWORK' Along with the change in scope there has been an organisation change. The journal's Editorial office has moved to Edinburgh from the publishers, Institute of Physics Publishing (IOPP), in Bristol, UK. Following the spirit of the closer Editorial involvement with the journal that this change implies, members of the Editorial Board and I welcome at all times comments and suggestions from authors, readers and referees that will help us to improve the journal according to the needs of the community we serve. Abstracting and Indexing `NETWORK is covered by the leading indexing services including Science Citation Index, Neuroscience Citation Index, Current Contents and Index Medicus/MEDLINE. `NETWORK' and Electronic Publishing IOPP has extensive experience in electronic publishing. Here are some of the current and future benefits of electronic publishing for authors and readers of `NETWORK': Papers that are submitted electronically can appear on the website very soon after they have been accepted, which can be at least a month before the printed journal appears. Electronic versions of papers can include full colour figures, and multimedia attachments can be provided free of charge; see www.iop.org/Journals/nfa for further details. IOPP Electronic Journals were accessed over 2.5 million times in 1998, of which 22,000 hits were aimed at `NETWORK' articles. As from 1st January 2000, readers of all IOPP journals will have access to the last 10 years on-line; this means that the whole of `NETWORK' will be accessible in this way. In the longer term, IOPP is developing a web-based system of refereeing that should lead to even faster publication times. In Conclusion For full information on the changes at `NETWORK', an Editorial will feature in the next issue of the journal. We are confident that the recent changes can only benefit authors and readers of `NETWORK. I am looking forward to establishing the journal as a premier European and world-wide contributor to the field of computational neuroscience. Professor David Willshaw Editor-in-Chief, `NETWORK: Computation in Neural Systems' Institute for Adaptive and Neural Computation Division of Informatics University of Edinburgh 2 Buccleuch Place Edinburgh EH8 9LW Scotland, UK Tel: +44 (0)131 650 4404/5 Fax: +44 (0)131 650 4406 Email: neted at anc.ed.ac.uk From cmbishop at microsoft.com Fri Oct 8 10:14:33 1999 From: cmbishop at microsoft.com (Christopher Bishop) Date: Fri, 8 Oct 1999 07:14:33 -0700 Subject: Postdoctoral Research Fellowship at Cambridge Message-ID: <3FF8121C9B6DD111812100805F31FC0D101F2650@RED-MSG-59> DARWIN COLLEGE, CAMBRIDGE Microsoft Research Fellowship http://www.dar.cam.ac.uk/ http://research.microsoft.com/cambridge Closing date: 15 October 1999. The Governing Body of Darwin College Cambridge, and Microsoft Research Limited, jointly invite applications for a stipendiary Research Fellowship supporting research in the field of adaptive computing (including topics such as pattern recognition, probabilistic inference, statistical learning theory and computer vision). Eligibility Men and women graduates of any university are eligible to apply, irrespective of age, provided they have a doctorate or an equivalent qualification, or expect to have submitted their thesis before taking up the Fellowship. Tenure The Fellowship will be tenable for two years commencing l October 2000 or on a date to be agreed. The successful candidate will engage in research full-time at the Microsoft Research Laboratory in Cambridge, and will join the new Machine Learning and Perception research group currently comprising Chris Bishop, Andrew Blake, Ben Bradshaw, Bernhard Schoelkopf, Michael Tipping and Phil Torr. The Fellow will be a member of the Governing Body of Darwin College and will be subject to the Statutes and Ordinances of the College which may be seen on request to the Bursar. The Statutes include the obligation to reside in or near Cambridge for at least two-thirds of each University term, but the Governing Body will normally excuse absences made necessary by the nature of the research undertaken. Stipend and Emoluments The stipend will be dependent upon age and experience. In addition the Fellow will be able to take seven meals per week at the College table free of charge and additional meals at his or her own expense. Guests may be invited to all meals (within the limits of available accommodation), ten of them free of charge within any quarter of the year. College accommodation will be provided, subject to availability, or an accommodation allowance will be paid in lieu. In addition to a salary the Fellowship provides funding for conference participation. Applications Applications should reach the Master, Darwin College, Cambridge CB3 9EU by 15 October 1999. They should be typed and should include SIX copies of (1) a curriculum vitae, (2) an account, in not more than 1000 words, of the proposed research, including a brief statement of the aims and background to it, (3) the names and addresses of three referees (including telephone, fax and e-mail co-ordinates), WHO SHOULD BE ASKED TO WRITE AT ONCE DIRECT TO THE COLLEGE, and (4) a list of published or unpublished work that would be available for submission if requested. Testimonials should not be sent. Electronic applications (in postscript, pdf, Word or text formats) may be sent to cmbishop at microsoft.com to arrive by 15 October 1999. Short-listed candidates may be asked to make themselves available for interview at Darwin College on a date to be arranged before mid-January 2000: election will be made as soon as possible thereafter. In certain circumstances travelling expenses for overseas interviewees may be covered. The College follows an equal opportunities policy. From dagli at umr.edu Sat Oct 9 15:31:33 1999 From: dagli at umr.edu (Cihan Dagli) Date: Sat, 9 Oct 1999 14:31:33 -0500 Subject: Artificial Neural Networks in Engineering Conference ANNIE'99 Message-ID: Dear Colleagues On behalf of the organizing committee I would like to invite you to attend ANNIE 1999, an international conference to be held on November 7-10, 1999, at Marriott's Pavilion Hotel in downtown, St. Louis, Missouri, USA. This will be the ninth international gathering of researchers interested in Smart Engineering System Design using neural networks, fuzzy logic, evolutionary programming, data mining, and complex systems. The previous conferences each drew approximately 150 papers from twenty countries with their proceedings published by ASME Press as hardbound books in eight volumes. The last volume, edited by Dagli, et. al., was titled ``Smart Engineering Systems: Neural Networks, Fuzzy Logic, Evolutionary Programming, Data Mining, and Rough Sets.'' Enclosed Please find the preliminary program for the conference. You can get the latest information about the conference at http://www.umr.edu/~annie/annie99/ and for the details of advanced registration to the conference at http://www.umr.edu/~annie/rf.htm . The conference will cover the theory of Smart Engineering System Design techniques, namely; neural networks, fuzzy logic, evolutionary programming, data mining, and complex systems. Presentations dealing with applications of these technologies are encouraged in the areas of: manufacturing engineering, biology and medicine, pattern recognition, image processing, process monitoring, control, recent theoretical developments in neural networks, fuzzy logic, data mining, rough sets, evolutionary programming, fractals, chaos, and wavelets that can impact Smart Engineering System Design. The response to ANNIE '99 was excellent with 310 abstracts received. Papers submitted based on these abstracts were reviewed by two referees and all accepted papers are included in the conference proceedings to be published by ASME Press as a hardbound book titled Smart Engineering System Design: Neural Networks, Fuzzy Logic, Evolutionary Programming, Data Mining and Complex Systems, edited by Drs. Dagli, Buczak, Ghosh, Embrechts and Ersoy. Conference banquet and the presentation of Best Paper awards are scheduled on Tuesday evening. Our Banquet Plenary Speaker for this year's ANNIE is University Professor Simon Haykin. He is the director of Neurocomputation for Signal Processing Group at McMaster University, Hamilton, Ontario, Canada. His banquet talk is titled "Neural Networks for Signal Processing." There are nine plenary sessions scheduled for ANNIE '99. Dr. Walter Freeman will present "The Neurodynamics of Intentionality is the Basis of Intelligent Behavior" at the Monday morning plenary session. At noon on Monday, Dr. Hojjat Adeli will present "Distributed and Neurocomputing for Large-Scale Engineering Design Automation." Monday afternoon Dr. Bhavani Thuraisingham will present " Data Mining and Knowledge Discovery: Developments and Challenges". In Tuesday plenary session, " Evolvable Hardware" will be presented by Dr. Adrian Stoica. Tuesday's luncheon plenary "Intangled Behavior and Bizarre Systems" will be presented by Dr. Steve Kercel. The Tuesday afternoon plenary will be "Neural Mechanics of the Cerebellum and Its Application" presented by Dr. Witali Dunin-Barkowski. Dr. Kristin Bennett will present "Support Vector Machines: An Overview and New Extensions" for the Wednesday morning plenary. The luncheon plenary " Searching the Web-it's worse than you think!" will be presented by Dr. Lee Giles. The Wednesday afternoon plenary will be "Neural Networks for Analysis of Data with Gaps" presented by Dr. Alexander Gorban. Half-day tutorials have been scheduled for Sunday, November 7th. These state-of-the-art workshops cover the following areas: Collective Intelligence Chaos Data Mining: Technologies, Tools and Trends Hybrid Intelligent Systems Using Soft Computing Techniques and Fractal Theory Computer Vision and Fuzzy-Neural Network Models Classifiers Ensembles: how and why they work Interior Point Optimization Methods in Support Vector Machines Training Novel Applications of Neural Network, Signal Processing and Data Mining Technologies An Introduction to Evolutionary Engineering Pattern Recognition Applications Development: The Process I would like to thank members of the Organizing Committee, Sponsoring Organization, and Co-Chairs, Drs. Buczak, Ghosh, Embrechts and Ersoy, for putting together an excellent program for ANNIE '99. I would like to recognize the excellent and timely efforts of the referees and contributors that made the conference possible. We are looking forward to your visit and participation at the meeting. Sincerely, Cihan H. Dagli Conference Chairman Smart Engineering Systems Laboratory 229 Engineering Management 1870 Miner Circle University of Missouri-Rolla Rolla MO 65409-0370 Phone: 573 341-4374 Fax: 573 341-6567 USA http://www.umr.edu/~annie http://www.gbhap.com/Smart_Engineering_System_Design/ http://www.umr.edu/~dagli From cg at santafe.edu Fri Oct 8 18:23:28 1999 From: cg at santafe.edu (Christine Gonzales) Date: Fri, 8 Oct 1999 16:23:28 -0600 Subject: Complex Systems Summer School, 2000 Message-ID: Santa Fe Institute Complex Systems Summer School, 2000 June 4 to June 30, 2000 in Santa Fe, New Mexico USA An intensive four-week introduction to complex behavior in mathematical, physical and biological systems, intended for graduate students and postdoctoral fellows. No tuition is charged. The first week will consist of an intensive series of lectures, demonstrations, and experiments introducing some core ideas and tools of complex systems research. The topics will include experimental and theoretical nonlinear dynamics and pattern formation, statistical mechanics and stochastic processes, information theory and computation theory, adaptive computation, and computer modeling tools. (Prior to week one, there may be an optional intensive two-day math review.) Weeks two and three will consist of short courses on current research in complex systems, and students will work on individual or team projects. Week four will be devoted to completion and presentation of student projects. Co-Directors: Raymond E. Goldstein, Physics and Applied Mathematics, University of Arizona; Melanie Mitchell, Biophysics, Los Alamos National Laboratory; and Lynn Nadel, Psychology, University of Arizona. Location: Held on the campus of the College of Santa Fe. Administered by the Santa Fe Institute. Application instructions: Provide a current resume with publications list (if available), statement of current research interests, comments about why you want to attend the school, and two letters of recommendation from scientists who know your work. Include your e-mail address and fax number. Send only complete application packages by postal mail to: Summer School, Santa Fe Institute 1399 Hyde Park Road Santa Fe, New Mexico USA 87501 505-984-8800 ext. 235 (v); 505-982-0565 (fax) February 7, 2000 deadline. Women and minorities encouraged to apply. Further information at http://www.santafe.edu/sfi/education/indexCSSS.html or summerschool at santafe.edu Christine Gonzales (505) 984-8800 ext. 235 Fax. 982-0565 From vaina at engc.bu.edu Mon Oct 11 15:29:11 1999 From: vaina at engc.bu.edu (Lucia M. Vaina) Date: Mon, 11 Oct 1999 15:29:11 -0400 Subject: Research position in fMRI applied to neurorecovery after stroke Message-ID: A non-text attachment was scrubbed... Name: not available Type: text/enriched Size: 2167 bytes Desc: not available Url : https://mailman.srv.cs.cmu.edu/mailman/private/connectionists/attachments/00000000/d562bf7d/attachment-0001.bin From ingber at ingber.com Mon Oct 11 22:28:30 1999 From: ingber at ingber.com (Lester Ingber) Date: Mon, 11 Oct 1999 21:28:30 -0500 Subject: EEG data available Message-ID: <19991011212830.A27943@ingber.com> I have had several requests for EEG data since I had to remove access from a large dataset I used for some of my studies, due to a disk crash at an archive at U Oregon. I am making a small dataset (4 MB) and two larger datasets (40 MB each) available at my Caltech archive. I do not know for how long I will be able to keep the larger datasets available. I cannot keep the largest dataset (700 MB) available there and do not plan on uploading it anyplace else. All files sizes are in .gz-compressed formats. See http://www.ingber.com/smni_eeg_data.html I hope you find this data useful for your own studies. Lester -- Lester Ingber http://www.ingber.com/ PO Box 06440 Wacker Dr PO Sears Tower Chicago IL 60606-0440 http://www.alumni.caltech.edu/~ingber/ From melnik at data.cs.brandeis.edu Tue Oct 12 12:55:41 1999 From: melnik at data.cs.brandeis.edu (Ofer Melnik) Date: Tue, 12 Oct 1999 12:55:41 -0400 (EDT) Subject: Network analysis page Message-ID: I'ld like to invite you to check out the webpage we put together to illustrate the Decision Intersection Boundary Algorithm (DIBA). The DIBA algorithm is used to analyze feed-forward threshold neural networks, by generating polytopics decision regions that describe network function. It makes a great pedagogical tool, clearly illustrating what neural networks do and how they learn. The page has multiple examples of networks analyzed with the algorithm, MOVIES of networks learning, some theoretical results and a link to the Tech-Report. The URL is: http://www.demo.cs.brandeis.edu/pr/DIBA/ -Ofer ----------------------------------------------------------------- Ofer Melnik melnik at cs.brandeis.edu Volen Center for Complex Systems Ph: (781)-736-2719 Brandeis University (781)-736-DEMO Waltham MA 02254 From pascal at icsc.ab.ca Tue Oct 12 12:27:27 1999 From: pascal at icsc.ab.ca (pascal@icsc.ab.ca) Date: Tue, 12 Oct 1999 10:27:27 -0600 Subject: Neural Computation Symposium Message-ID: <19991012162730Z25908-276+18008@mail.compusmart.ab.ca> ANNOUNCEMENT / CALL FOR PAPERS Second International ICSC Symposium on NEURAL COMPUTATION / NC'2000 To be held at the Technical University of Berlin, Germany May 23-26, 2000 http//www.icsc.ab.ca/nc2000.htm SYMPOSIUM CHAIR Prof. Hans Heinrich Bothe Technical University of Denmark (DTU) Department of Information Technology Building 344 DK-2800 Lyngby, Denmark Email: hhb at it.dtu.dk Phone: +45-4525-3632 Fax: +45-4588-0117 PUBLICATION CHAIR Prof. Raul Rojas Freie Universit?t Berlin Institut Informatik / FB Mathematik Takustrasse 9 D - 14195 Berlin / Germany Email: rojas at inf.fu-berlin.de Fax: +49-30-8387-5109 SYMPOSIUM ORGANIZER ICSC International Computer Science Conventions P.O. Box 279 Millet, Alberta T0C 1Z0 / Canada Phone: +1-780-387-3546 Fax: +1-780-387-4329 Email: operating at icsc.ab.ca WWW: http//www.icsc.ab.ca INTERNATIONAL PROGRAM COMMITTEE Igor Aleksander, Imperial College of Science & Technology, London, U.K. Peter G. Anderson, Rochester Institute of Technology, NY, USA Horst Bischof, Technical University Vienna, Austria Ruediger W. Brause, J.W. Goethe-University, Frankfurt, Germany Juan Lopez Coronado, Universidad Polotecnica de Cartagena, Spain Ludwig Cromme, Brandenburgische Technische Universitaet Cottbus, Germany Chris deSilva, University of Western Australia Crawley, Australia Georg Dorffner, Austrian Research Institute for Artificial Intelligence, Vienna, Austria Gunhan Dundar, Bogazici University, Istanbul, Turkey Bernd Fritzke, Dresden University of Technology, Dresden, Germany Kunihiko Fukushima, The University of Electro-Communications, Tokyo, Japan Wulfram Gerstner, EPFL Lausanne, Switzerland Stan Gielen, University of Nijmegen, Netherlands Marco Gori, University of Siena, Italy Bruce Graham, University of Edinburgh, U.K. Dorothea Heiss-Czedik, Technical University Vienna, Austria Michael Heiss, Siemens Vienna, Austria Giacomo Indiveri, Swiss Federal Institute of Technology, Zurich, Switzerland Lakhmi C. Jain, University of South Australia, The Levels, Australia Nikola Kasabov, University of Otago, Dunedin, New Zealand Bart Kosko, University of Southern California, Los Angeles CA, USA Rudolf Kruse, University of Magedburg, Germany Te-Won Lee, Salk Institute & University of California, San Diego, USA Fa-Long Luo, R&D Department, Redwood City, USA G.Nicolas Marichal, University of La Laguna Tenerife, Spain Giuseppe Martinelli, University of Rome 1, Italy Klaus-Robert Mueller, GMD First, Berlin, Germany Fazel Naghdy, University of Wollongong, Australia M. Palaniswami, University of Melbourne, Australia Guenther Palm, University of Ulm, Germany Yoh-Han Pao, Case Western Reserve University, Cleveland OH, USA Alexander V. Pavlov, Laboratory for Optical Fuzzy Systems, St. Petersburg, Russia Witold Pedrycz, University of Alberta, Edmonton, Canada Raul Rojas, Freie Universit?t Berlin, Germany V. David Sanchez A., Falon, Inc., San Diego CA, USA Bernd Schuermann, Siemens ZFE, Munich, Germany J.S. Shawe-Taylor, Royal Holloway University of London, U.K. Peter Sincak, Technical University of Kosice, Slovakia Nigel Steele, Coventry University, U.K. Rainer Stotzka, Forschungszentrum Karlsruhe, Germany Piotr Szczepaniak, Technical University of Lodz, Poland Csaba Szepesvari, University of Szeged, Hungary Henning Tolle, Technische Hochschule Darmstadt, Germany Shiro Usui, Toyohashi University of Technology, Toyohashi, Japan Luis Alfredo Vidal de Carvalho, Federal University of Rio de Janeiro, Brazil Roberto C. Villas-Boas, United Nations Industrial Development Organization, Rio de Janeiro, Brazil Lipo Wang, Nanyang Technological University, Singapore Klaus Weber, Technical University of Cottbus, Germany Andreas Weingessel, Technical University Vienna, Austria Takeshi Yamakawa, Kyushu Institute of Technology, Fukuoka, Japan Andreas Zell, Universitaet Tuebingen, Germany Tom Ziemke, University of Skoevde, Sweden Jacek M. Zurada, University of Louisville, K.Y., USA ************************************************* SPONSORS Technical University of Berlin, Germany Fraunhofer-Gesellschaft fuer Produktionsanlagen, Berlin, Germany (FHG-IPK) ICSC International Computer Science Conventions, Canada/Switzerland ************************************************* INTRODUCTION The science of neural computation focusses on mathematical aspects to solve complex practical problems, and it also seeks to help neurology, brain theory and cognitive psychology in the understanding of the functioning of the nervous system by means of computational models of neurons, neural nets and subcellular processes. NC'2000 aims to become a major point of contact for research scientists, engineers and practitioners throughout the world in the field of Neural Computation. Participants will share the latest research, developments and ideas in the wide arena of disciplines encompassed under the heading of NC'2000 as a follow-up of the most successful NC'98 conference in Vienna, Austria. ************************************************* TOPICS Contributions are sought in areas based on the list below, which is indicative only. Contributions from new applications areas are welcome. COMPUTATIONAL NEURAL NETWORK MODELS - Artificial neural network paradigms - Knowledge representation - Learing and generalization - Probabilistic neural networks - Information theoretic approaches - Time-coded neural networks - Pulse-coded neural networks - Self-organization - Cellular automata - Hybrid systems (e.g. neuro-fuzzy, GA, evolutionary strategies) - Chaos in neural networks - Statistical and numerical aspects NEUROPHYSIOLOGICALLY INSPRIED MODELS - Neurophysiological foundations - Spiking neuron models and neuron assemblies - Models of brain centers and sensory pathways - Sensormotor integration - Sensation, Perception and Attention - Spatio-temporal Orientation - Reactive Behavior SOFTWARE AND HARDWARE IMPLEMENTATIONS - Simulation and Graphical Programming Tools - Distributed Systems - Neuro-chips, -controllers and -computers - Analog and Digital Electronic Implementations - Optic, Holographic Implementations NEURAL NETWORK APPLICATIONS - Pre-processing and Feature Extraction - Sound, Speech and Image Processing - Pattern Recognition and System Identification - Computer Vision, Feature Binding and Image Understanding - Autonomous Sensor Systems, Multivariate Sensor Fusion - Robotics and Control - Behavior based Exploration and Planning - Power Systems - Environmental Systems - Decision Support Systems - Medical Applications - Operational Research and Logistics ************************************************* SCIENTIFIC PROGRAM NC'2000 will include invited plenary talks, contributed sessions, invited sessions, workhops and tutorials. At present, the following special sessions are in preparation SELF-ORGANIZING MAPS MODIFICATION, EXTENSION AND APPLICATION with a tutorial by Prof. Teuvo Kohonen, who is the 'inventor' of SOM's Session Chair Dr. Udo Seiffert, University of Magdeburg, Germany NONLINEAR DYNAMICS AND NEURAL FIELDS FOR THE ANALYSIS AND ORGANIZATION OF BEHAVIOR with participation by Prof. S.-I. Amari, Riken Institute, Japan, who is the founder of the mathematical formulation of neural fields. Session Chair Prof. Werner von Seelen, Ruhr-University Bochum, Germany ANN IN BIOMEDICINE Session Chair Prof. Piotr Szczepaniak, Technical University of Lodz, Poland OPTICS IN FUZZY SET THEORY Session Chair Dr. Alexander V. Pavlov, Laboratory for Optical Fuzzy Systems, St. Petersburg, Russia ************************************************* INVITED SESSIONS The organization of invited sessions is encouraged. Prospective organizers are requested to send a session proposal (consisting of 4-5 invited papers, the recommended session-chair and co-chair, as well as a short statement describing the title and the purpose of the session to the Symposium Chairman or the Symposium Organizer. Invited sessions should preferably start with a tutorial paper. The registration fee of the session organizer will be waived, if at least 4 authors of invited papers register to the conference. ************************************************* POSTER PRESENTATIONS Poster presentations are encouraged for people who wish to receive peer feedback and practical examples of applied research are particularly welcome. Poster sessions will allow the presentation and discussion of respective papers, which will also be included in the conference proceedings. ************************************************* WORKSHOPS, TUTORIALS AND OTHER CONTRIBUTIONS Proposals should be submitted as soon as possible to the Symposium Chairman or the Symposium Organizer. ************************************************* SUBMISSION OF PAPERS Prospective authors are requested to either send a draft paper (maximum 7 pages) or an extended abstract for review by the International Program Committee. All papers must be written in English, starting with a succinct statement of the problem, the results achieved, their significance and a comparison with previous work. Submissions must be received by October 31, 1999. Regular papers, as well as poster presentations, tutorial papers and invited sessions are encouraged. The submission should also include - Title of conference (NC'2000) - Type of paper (regular, poster, tutorial or invited) - Title of proposed paper - Authors names, affiliations, addresses - Name of author to contact for correspondence - E-mail address and fax # of contact author - Topics which best describe the paper (max. 5 keywords) - Short CV of authors (recommended) Contributions are welcome from those working in industry and having experience in the topics of this conference as well as from academics. The conference language is English. It is strongly recommended to submit abstracts by electronic mail to operating at icsc.ab.ca or else by fax or mail (2 copies) to the following address ICSC Canada P.O. Box 279 Millet, Alberta T0C 1Z0 Canada Fax +1-780-387-4329 ************************************************* BEST PRESENTATION AWARDS The best oral and poster presentations will be honored with best presentation awards. ************************************************* PUBLICATIONS Conference proceedings (including all accepted papers) will be published by ICSC Academic Press and be available for the delegates at the symposium in printed form or on CD-ROM. Authors of a selected number of innovative papers will be invited to submit extended manuscripts for publication in prestigious international journals. ************************************************* IMPORTANT DATES - Submission Deadline: October 31, 1999 - Notification of Acceptance: December 31, 1999 - Delivery of full papers: February 15, 2000 - Tutorials and Workshops: May 23, 2000 - NC'2000 Symposium: May 24-26, 2000 ************************************************* ACCOMMODATION Accommodation at reasonable rates will be available at nearby hotels. Full details will follow with the letters of acceptance. ************************************************* SOCIAL AND TOURIST ACTIVITIES A social program will be organized and also be available for accompanying persons. ************************************************* THE FASCINATING CITY OF BERLIN The old and new capital of Germany is a mecca for scientists and cultural enthusiasts, for day workers and night owls. Charming with its several opera houses, concert halls, cabarets, and beer gardens, Berlin is full of spontaneous cultural events or happenings. As an urban building site for the future, it is at the same time a living contradiction and a casual place with relaxation areas and large parks right in the city center. The fine nature, pine tree forests, and 1001 lakes around the city supply Berlin with it s very specific sparkling air in spring time. No other city in Germany has during the last 100 years played such a prominent role in history and in the imagination of the people Social and industrial revolution in the last century, world war I and later manifestation of the first German republic, the 'Golden Twenties', nazis dictatorship and world war II, splitting by the Berlin Wall, 'economic miracle' in the west and socialistic showpiece city in the east, '68 stu dent and alternative lifestyle movement in the west and peace movement in the east, and finally, the fall of the Wall. Berlin tempts with its many research facilities, and it is Germany's largest industrial city with headquarters or dependences of most major companies. At present, approximately 3.5 million inhabitants live in the reunified city, among which more than 120.000 students, who study in three universities and twelve colleges or schools of arts. Berlin ist eine Reise wert. Welcome! ************************************************* FURTHER INFORMATION Fully updated information is available from http//www.icsc.ab.ca/nc2000.htm You may also contact - ICSC International Computer Science Conventions P.O. Box 279 Millet, Alberta T0C 1Z0 / Canada Phone: +1-780-387-3546 Fax: +1-780-387-4329 Email: operating at icsc.ab.ca or, for specific scientific requests, the symposium chairman. From pcohen at cse.ogi.edu Tue Oct 12 17:44:52 1999 From: pcohen at cse.ogi.edu (Phil Cohen) Date: Tue, 12 Oct 1999 14:44:52 -0700 Subject: Postdoctoral Position: Gesture recognition at OGI Message-ID: <007601bf14fb$03f41d60$78365f81@cse.ogi.edu> The following position is available at the Center for Human-Computer Communication at the Oregon Graduate Institute: TOPIC: Two-dimensional and three-dimensional gesture recognition as part of multimodal systems. You will be an integral member of a multidisciplinary team that has developed leading edge technology. Topics will include gesture interaction for 2-D map-based interactions, and also for 3-D real and virtual worlds. EXPERIENCE: Postdoctoral researcher (new Ph.D. or more experienced), extensive knowledge of neural networks (especially committee-based approaches), and HMMs. Prior work with gesture recognition and/or handwriting recognition is essential. Experience with natural language processing would also be desirable. SOFTWARE DEVELOPMENT ENVIRONMENTS: Should be familiar with MatLab, and program in C++ on PC's and Unix. Experience with Windows CE would also be desirable. Compensation will be competitive. INSTITUTION: CHCC at OGI is one of the world's leading institutions engaged in research on multimodal interaction. We are well-supported by DARPA, NSF, ONR, and other organizations. The QuickSet multimodal system is highly regarded and is a key part of numerous Government programs. To learn more about our Center and our work, please visit http://www.cse.ogi.edu/CHCC. AREA: Portland, Oregon is a fine place to live. There are numerous outdoors activities (skiing, windsurfing, hiking, beach combing, etc.) within 1 hour's drive. Yes, it rains here in winter (but it probably does something nasty where you live too.) However, one can often have 5 months of no/little rain in Spring/Summer/Fall. To apply, please send a resume, research statment, and names/contact information for at least 3 references to: Prof. Phil Cohen Center for Human-Computer Communication Oregon Graduate Institute of Science and Technology 20000 NW Walker Rd. Beaverton, OR 97006 503-690-1326 Fax: 503-690-1548 From Ronan.Reilly at ucd.ie Wed Oct 13 18:43:43 1999 From: Ronan.Reilly at ucd.ie (Ronan Reilly) Date: Wed, 13 Oct 1999 23:43:43 +0100 Subject: UCD, Ireland - Two-year graduate studentship Message-ID: <02aa01bf15cc$6a0de370$5f042b89@ucd.ie> Two-year graduate studentship on a computational neuroscience project Department of Computer Science, University College Dublin, Ireland The Cortical Software Re-Use (CSRU) project is a two-year interdisciplinary project involving the Departments of Mathematical Physics, Psychology, and Computer Science. Its aim is to apply a dynamical systems approach to modelling cortical computation with a specific focus on frontal lobe function. The Department of Computer Science at University College Dublin has funding of ?12,700 per year for two years for an MSc or PhD student to work on the project. The ideal candidate will have a background in computer science and some experience of computational modelling in the neuroscience area. For further details, contact Ronan.Reilly at ucd.ie. ________________________________ Ronan G. Reilly, PhD Department of Computer Science University College Dublin Belfield Dublin 4 Ireland voice: +353-1-706 2475 fax: +353-1-269 7262 e-mail: Ronan.Reilly at ucd.ie web: http://www.cs.ucd.ie/staff/rreilly/ From gbugmann at soc.plym.ac.uk Wed Oct 13 23:21:00 1999 From: gbugmann at soc.plym.ac.uk (Guido.Bugmann xtn 2566) Date: Thu, 14 Oct 1999 04:21:00 +0100 (BST) Subject: Postdoctoral Research Associate and one PhD student Message-ID: Dear Moderator, the annoucement below may be of interest to members of the list, the project involves learning (via instruction) and knowledge representation (probably connectionnist). Thanks in advance for posting this announcement. Guido Bugmann ----------------------------- Dr. Guido Bugmann Senior Research Fellow Centre for Neural and Adaptive Systems School of Computing University of Plymouth Plymouth PL4 8AA United Kingdom ----------------------------- Tel: (+44) 1752 23 25 66 / 41 Fax: (+44) 1752 23 25 40 Email: gbugmann at soc.plym.ac.uk or gbugmann at plymouth.ac.uk http://www.tech.plym.ac.uk/soc/Staff/GuidBugm/Bugmann.htm ----------------------------- Sorry if you receive multiple copies of this announcements. ----------------------------------------------------------- A Postdoctoral Research Associate and one PhD student are sought for an exciting project aimed at developing principles of "Instruction-Based Learning", i.e. teaching robots using natural language. This project explores a still-untapped method of knowledge acquisition and learning by intelligent systems: the acquisition of knowledge from Natural Language (NL) instruction. This is very effective in human learning and will be essential for adapting future intelligent systems to the needs of naive users. The aim of the project is to investigate real world Instruction Based-Learning (IBL) in a generic route instruction task. Users will engage in a dialogue with a mobile robot equipped with artificial vision, in order to teach it how to navigate a simplified maze-like environment. This experimental set-up will limit perceptual and control problems and also reduce the complexity of NL processing. The research will focus on the problem of how NL instructions can be used by an intelligent embodied agent to build a hierarchy of complex functions based on a limited set of low-level perceptual, motor and cognitive functions. We will investigate how the internal representations required for robot sensing and navigation can support a usable speech-based interface. Given the use of artificial vision and voice input, such a system can contribute to assisting visually impaired people and wheelchair users. This 3 year project for 2 research associates and one PhD student is funded by the UK Engineering and Physical Sciences Research Council. It is jointly managed by Dr. G. Bugmann of the Centre for Neural and Adaptive Systems of the University of Plymouth and by Dr. E. Klein of the Institute for Communicating and Collaborative Systems at the University of Edinburgh. The tasks are subdivided into two broad categories: (i) natural language processing and dialogue management at Edinburgh (1 RA) and (ii) robot experiments, knowledge representation and systems integration in Plymouth (1 RA, 1 RS). The project will involve sustained interaction and coordination between the two groups in order to support the required level of integration. The positions offered in Plymouth are: 1. Research Associate This position requires a strong background in the areas of systems architecture, knowledge representation, systems integration, robotics and artificial vision. Salary scale: 15205-19999 UK Pounds (Pay award pending.) 2. Research student The PhD student will examine some of the more speculative and open-ended aspects of IBL as a form of robot learning. Bursary: 6805 UK Pounds (funding for non-UK students submitted to restrictions). The two positions require skills in programming C++ and Perl/Python under Windows NT or LINUX. Experience with hardware issues and real world image processing is desirable. Applications should be received before 1 December, but the positions will remain open until suitable candidates are found. Candidates should send a CV, a description of motivations and the names of three references (by surface mail of email) to: Carole Watson School of Computing University of Plymouth Drake Circus Plymouth PL4 8AA, UK c.watson at plymouth.ac.uk Tel: +44 (0) 1752 23 25 41 Fax: +44 (0) 1752 23 25 40 Informal questions can be asked to: Dr. Guido Bugmann (abroad from 8 - 18 Oct.) gbugmann at soc.plym.ac.uk Tel: +44 (0) 1752 23 25 66 Fax: +44 (0) 1752 23 25 40 http://techweb.plym.ac.uk/soc/staff/guidbugm/bugmann.htm ----------------------------- From svensen at cns.mpg.de Thu Oct 14 03:11:09 1999 From: svensen at cns.mpg.de (Markus Svensen) Date: Thu, 14 Oct 1999 09:11:09 +0200 Subject: Post-doc, Medical Image Processing Message-ID: <3805820D.677F190A@cns.mpg.de> Dear Connectionists, please do not respond to me, but to Dr Frithjof Kruggel (details below). Markus Svensen RESEARCH POSITION Max-Planck-Institute of Cognitive Neuroscience Workgroup on Signal and Image Processing The Workgroup on Signal and Image Processing is seeking applicants for a Post-Doctoral Research Position in the area of image processing as applied to the analysis of anatomical data as revealed by magnetic resonance imaging (MRI). The successful applicant will be expected to participate in an advanced research and development effort centered on the following topics: o image segmentation of pathological MRI data o statistical analysis of characteristics of brain lesions o detection of changes in brain lesions with time Successful applicants should have an advanced degree in Mathematics, Physics, Computer Science or Medical Informatics, and significant experience in image processing as applied to medical problems. The term of the position is currently limited to one year, with a possible extension of another year. A starting date around January 2000 is expected. A competitive salary will be offered, with actual salary dependent on the applicant's qualifications and experience. We encourage qualified women to apply. Interested applicants should send a resume and a list of references to Dr. F. Kruggel (kruggel at cns.mpg.de) before October 31, 1999. Electronic application is preferred, but if necessary, applications may be sent to the address below. Dr. F. Kruggel Max-Planck-Institute of Cognitive Neuroscience Stephanstr. 1, 04103 Leipzig, Germany kruggel at cns.mpg.de Phone: (49) 341 9940 223, FAX: (49) 341 9940 221 From aweigend at stern.nyu.edu Thu Oct 14 07:10:03 1999 From: aweigend at stern.nyu.edu (Andreas Weigend) Date: Thu, 14 Oct 1999 07:10:03 -0400 (EDT) Subject: Jobs: Music Cognition / Machine Learning / Web Message-ID: <199910141110.HAA05655@sabai.stern.nyu.edu> A non-text attachment was scrubbed... Name: not available Type: text Size: 1359 bytes Desc: not available Url : https://mailman.srv.cs.cmu.edu/mailman/private/connectionists/attachments/00000000/16047f82/attachment-0001.ksh From jagota at cse.ucsc.edu Thu Oct 14 14:04:16 1999 From: jagota at cse.ucsc.edu (Arun Jagota) Date: Thu, 14 Oct 1999 11:04:16 -0700 Subject: new e-survey in NCS Message-ID: <199910141804.LAA10346@sundance.cse.ucsc.edu> New refereed e-publication action editor: Yoshua Bengio W. Duch & N. Jankowski, Survey of Neural Transfer Functions Neural Computing Surveys 2, 163--212, 1999. 91 references. http://www.icsi.berkeley.edu/~jagota/NCS Abstract: The choice of transfer functions may strongly influence complexity and performance of neural networks. Although sigmoidal transfer functions are the most common, there is no /a priori/ reason why models based on such functions should always provide optimal decision borders. A large number of alternative transfer functions have been described in the literature. A taxonomy of activation and output functions is proposed, and advantages of various non-local and local neural transfer functions are discussed. Several lesser known types of transfer functions, and new combinations of activation/output functions are described. Universal transfer functions, parametrized to change from localized to delocalized type, are of greatest interest. Other types of neural transfer functions discussed here include functions with activations based on non-Euclidean distance measures, bicentral functions, formed from products or linear combinations of pairs of sigmoids, and extensions of such functions making rotations of localized decision borders in highly dimensional spaces practical. Nonlinear input preprocessing techniques are briefly described, offering an alternative way to change the shapes of decision borders. From stefan.wermter at sunderland.ac.uk Thu Oct 14 14:23:21 1999 From: stefan.wermter at sunderland.ac.uk (Stefan Wermter) Date: Thu, 14 Oct 1999 19:23:21 +0100 Subject: Neural networks based on neuroscience Message-ID: <38061F99.5444E062@sunderland.ac.uk> In September 1999 we held an EPSRC funded EmerNet workshop, an International Workshop on Emergent Neural Computational Architectures Based on Neuroscience. Topics in the Call for Neural Networks based on Neuroscience had included: 1.Synchronisation: How does the brain synchronise its processing? How does the brain schedule its processing? 2.Processing speed: How does the brain compute with relatively slow computing elements but still achieve rapid and real time performance? 3.Robustness: How does human memory manage to continue to operate despite failure of its components? 4.Modular construction: What can we learn from the brain for building modular more powerful artificial neural network architectures to solve larger tasks? 5.Learning in context: How can we build learning algorithms which consider context? How can we design incremental learning algorithms and dynamic architectures? This message is to announce that a summary as well as all the papers of this workshohp have been made available at: http://www.his.sunderland.ac.uk/emernet/icann99w.html Also at http://www.his.sunderland.ac.uk/emernet/ you may find additional related information on this topic best wishes, Stefan Wermter *************************************** Professor Stefan Wermter Research Chair in Intelligent Systems University of Sunderland Centre of Informatics, SCET St Peters Way Sunderland SR6 0DD United Kingdom phone: +44 191 515 3279 fax: +44 191 515 2781 email: stefan.wermter at sunderland.ac.uk http://www.his.sunderland.ac.uk/~cs0stw/ http://www.his.sunderland.ac.uk/ **************************************** From mlf at dlsi.ua.es Thu Oct 14 02:56:57 1999 From: mlf at dlsi.ua.es (Mikel L. Forcada) Date: Thu, 14 Oct 1999 08:56:57 +0200 Subject: Paper available Message-ID: <38057EB9.18293B09@dlsi.ua.es> Dear connectionists: The following paper, to appear in Neural Computation is now available via our web server: PDF: http://www.dlsi.ua.es/~mlf/stanc3web.pdf gzipped PostScript: http://www.dlsi.ua.es/~mlf/stanc3web.ps.gz Stable encoding of finite-state machines in discrete-time recurrent neural nets with sigmoid units Rafael C. Carrasco, Mikel L. Forcada, M. ?ngeles Vald?s-Mu?oz Departament de Llenguatges i Sistemes Inform?tics Universitat d'Alacant E-03071 Alacant (Spain) Ram?n P. ?eco Departament de Ci?ncies Experimentals i Tecnologia Universitat "Miguel Hern?ndez" E-03202 Elx (Spain) ABSTRACT In recent years, there has been a lot of interest in the use of discrete-time recurrent neural nets (DTRNN) to learn finite-state tasks, with interesting results regarding the induction of simple finite-state machines from input--output strings. Parallel work has studied the computational power of DTRNN in connection with finite-state computation. This paper describes a simple strategy to devise stable encodings of finite-state machines in computationally capable discrete-time recurrent neural architectures with sigmoid units, and gives a detailed presentation on how this strategy may be applied to encode a general class of finite-state machines in a variety of commonly-used first- and second-order recurrent neural networks. Unlike previous work that either imposed some restrictions to state values, or used a detailed analysis based on fixed-point attractors, the present approach applies to any positive, bounded, strictly growing, continuous activation function, and uses simple bounding criteria based on a study of the conditions under which a proposed encoding scheme guarantees that the DTRNN is actually behaving as a finite-state machine. -- _____________________________________________________________________ Mikel L. Forcada E-mail: mlf at dlsi.ua.es Departament de Llenguatges Phone: +34-96-590-3400 ext. 3384; i Sistemes Informatics also +34-96-590-3772. UNIVERSITAT D'ALACANT Fax: +34-96-590-9326, -3464 E-03071 ALACANT, Spain. ______________________________________________________________________ URL: http://www.dlsi.ua.es/~mlf ---------------------------------------------------------------------- From jpittman at microsoft.com Thu Oct 14 16:16:26 1999 From: jpittman at microsoft.com (Jay Pittman) Date: Thu, 14 Oct 1999 13:16:26 -0700 Subject: job posting at Microsoft Message-ID: <783D93998201D311B0CF00805FEAA07B010DF872@RED-MSG-42> Here's a job posting that may be of interest to the readers of this list. The Microsoft Handwriting Recognition team has immediate openings in handwriting recognition research and development. Responsibilities center on incorporating research into the Microsoft Handwriting Recognition (HR) engine. Experience in handwriting recognition technology is not required, but a solid background in Computer Science, Mathematics, or Statistics, is needed, as is the desire and ability to learn new techniques from textbooks or research papers. The HR engine is being developed for English and Japanese and includes components that implement neural nets, hidden Markov models, signal processing, statistical search, n-gram statistical language modeling, and context-free grammar language modeling. Improvements to the HR engine are targeted at increasing performance and accuracy, and improving usability problems for mainstream use in applications, such as note taking into future versions of Microsoft Office. Qualifications required include at least a Bachelors degree in Computer Science, Mathematics, EE/Signal Processing, or related field, and strong C/C++ coding skills. You must be able to code up your ideas to improve the recognizer in C/C++. A strong interest and ability to write and support your own code in C/C++ is crucial to being successful in the group. Send your resume, complete with references (if that is practical), to: Dr. James A. Pittman jpittman at microsoft.com I prefer email in either Word or ascii format, but you can also snailmail to: Dr. James A. Pittman, 30/3092 Microsoft Corporation One Microsoft Way Redmond, WA 98052 From ingber at ingber.com Thu Oct 14 19:34:44 1999 From: ingber at ingber.com (Lester Ingber) Date: Thu, 14 Oct 1999 18:34:44 -0500 Subject: More EEG data available Message-ID: <19991014183444.A3457@ingber.com> Since my previous announcement below, I was contacted by Padhraic Smyth who invited my to upload this data to the KDD Archive [http://kdd.ics.uci.edu] hosted by the Department of Information and Computer Science, University of California, Irvine. Working with the Librarian of this archive, Stephen Bay, I uploaded the full data set. Access is available via http://www.ingber.com/smni_eeg_data.html or http://kdd.ics.uci.edu/databases/eeg/eeg.html Lester : I have had several requests for EEG data since I had to remove access : from a large dataset I used for some of my studies, due to a disk crash : at an archive at U Oregon. : : I am making a small dataset (4 MB) and two larger datasets (40 MB : each) available at my Caltech archive. I do not know for how long I : will be able to keep the larger datasets available. I cannot keep the : largest dataset (700 MB) available there and do not plan on uploading : it anyplace else. All files sizes are in .gz-compressed formats. See : http://www.ingber.com/smni_eeg_data.html : : I hope you find this data useful for your own studies. -- Lester Ingber http://www.ingber.com/ PO Box 06440 Wacker Dr PO Sears Tower Chicago IL 60606-0440 http://www.alumni.caltech.edu/~ingber/ From steve at cns.bu.edu Sat Oct 16 08:03:54 1999 From: steve at cns.bu.edu (Stephen Grossberg) Date: Sat, 16 Oct 1999 08:03:54 -0400 Subject: A Neural Model of Motion Processing and Visual Navigation by Cortical Area MST Message-ID: The following article can be read at http://cns-web.bu.edu/Profiles/Grossberg/ A Neural Model of Motion Processing and Visual Navigation by Cortical Area MST, by Stephen Grossberg, Ennio Mingolla, and Christopher Pack. Cerebral Cortex, in press Cells in the dorsal medial superior temporal cortex (MSTd) process optic flow generated by self-motion during visually-guided navigation. A neural model shows how interactions between well-known neural mechanisms (log polar cortical magnification, Gaussian motion-sensitive receptive fields, spatial pooling of motion-sensitive signals, and subtractive extraretinal eye movement signals) lead to emergent properties that quantitatively simulate neurophysiological data about MSTd cell properties and psychophysical data about human navigation. Model cells match MSTd neuron responses to optic flow stimuli placed in different parts of the visual field, including position invariance, tuning curves, preferred spiral directions, direction reversals, average response curves, and preferred locations for stimulus motion centers. The model shows how the preferred motion direction of the most active MSTd cells can explain human judgments of self-motion direction (heading), without using complex heading templates. The model explains when extraretinal eye movement signals are needed for accurate heading perception, and when retinal input is sufficient, and how heading judgments depend on scene layouts and rotation rates. From harnad at coglit.ecs.soton.ac.uk Sat Oct 16 13:29:37 1999 From: harnad at coglit.ecs.soton.ac.uk (Stevan Harnad) Date: Sat, 16 Oct 1999 18:29:37 +0100 (BST) Subject: LANGUAGE EVOLUTION: 2 Psycoloquy Calls for Commentators Message-ID: (1) THE COGNITIVE PREREQUISITES FOR LANGUAGE (Burling) http://www.cogsci.soton.ac.uk/psyc-bin/newpsy?10.032 (2) LANGUAGE EVOLUTION AND THE COMPLEXITY CRITERION (Bichakjian) http://www.cogsci.soton.ac.uk/psyc-bin/newpsy?10.033 The two target articles whose abstracts follow below were published today in PSYCOLOQUY, a refereed journal of Open Peer Commentary sponsored by the American Psychological Association. Qualified professional biobehavioral, neural or cognitive scientists are hereby invited to submit Open Peer Commentary on either or both articles. Please email or consult the journal's websites below for Instructions if you are not familiar with format or acceptance criteria for PSYCOLOQUY commentaries (all submissions are refereed). To submit articles and commentaries or to seek information: EMAIL: psyc at pucc.princeton.edu URL: http://www.princeton.edu/~harnad/psyc.html http://www.cogsci.soton.ac.uk/psyc ----------------------------------------------------------------------- (1) THE COGNITIVE PREREQUISITES FOR LANGUAGE Target Article on Language-Prerequisites Robbins Burling Department of Anthropology 1020 LSA Building University of Michigan Ann Arbor MI 48109 USA rburling at umich.edu ABSTRACT: The first use of words by our early ancestors probably depended on four cognitive capacities: A rich conceptual understanding of the world around us; the ability to use and understand motivated signs, both icons and indices; the ability to imitate; the ability to infer the referential intentions of others. The latter three capacities are rare or absent in nonprimate mammals, but incipient in apes and well developed in modern humans. Before early humans could have begun to use words these capacities would have needed further development than is found in modern apes. It is not clear why selection favoured these skills more strongly in our ancestors than in the ancestors of apes. KEYWORDS: cognition; evolution; iconicity; imitation; language; names; theory-of-mind; words. Access full text at: http://www.cogsci.soton.ac.uk/psyc-bin/newpsy?10.032 ----------------------------------------------------------------------- (2) LANGUAGE EVOLUTION AND THE COMPLEXITY CRITERION Target Article on Language-Complexity Bernard H. Bichakjian Department of French University of Nijmegen, The Netherlands Bichakjian at let.kun.nl http://welcome.to/bichakjian ABSTRACT: Though it is increasingly accepted in the behavioral sciences, the evolutionary approach is still meeting resistance in linguistics. Linguists generally cling to the idea that alternative linguistic features are simply gratuitous variants of one another, while the advocates of innate grammars, who make room for evolution as a biological process, exclude the evolution of languages. The rationale given is that today's languages are all complex systems. This argument is based on the failure to distinguish between complexities of form and function. The proper analysis reveals instead that linguistic features have consistently decreased their material complexity, while increasing their functionality. A systematic historical survey will show instead that languages have evolved and linguistic features have developed along a Darwinian line. KEYWORDS: complexity, Indo-European, language evolution, lateralization, neoteny, word order. Access full text at: http://www.cogsci.soton.ac.uk/psyc-bin/newpsy?10.033 ----------------------------------------------------------------------- To submit articles and commentaries or to seek information: EMAIL: psyc at pucc.princeton.edu URL: http://www.princeton.edu/~harnad/psyc.html http://www.cogsci.soton.ac.uk/psyc From bap at cs.unm.edu Sun Oct 17 03:47:24 1999 From: bap at cs.unm.edu (Barak Pearlmutter) Date: Sun, 17 Oct 1999 01:47:24 -0600 (MDT) Subject: Positions Available Message-ID: POSTDOCTORAL, PREDOCTORAL and STAFF POSITIONS AVAILABLE in Blind Source Separation and Magnetoencephalography Prof Barak A. Pearlmutter Prof Akaysha C. Tang Dept of Computer Science Dept of Psychology Dept of Neurosciences Dept of Neurosciences University of New Mexico University of New Mexico http://www.cs.unm.edu/~bap/ http://kongzi.unm.edu Our growing Brain and Computation group is focusing most of its effort on functional brain imaging, magnetoencephalography in particular. We are applying blind source separation techniques to MEG data, and designing and conducting experiments that take advantage of MEG's unique combination of high temporal resolution and spatial sensitivity. The project spans the spectrum from experiment design and data acquisition to analysis and theory, and we have positions available at all levels for neuroscientists, computer scientists, physicists, differential geometers, etc who can contribute to - relevant theoretical work - developing source separation algorithms - implementing algorithms for processing brain imaging data - developing visualization tools and techniques - contributing to our open-source MEG software suite - designing and conducting experiments - collecting and analyzing data These positions are to be largely supported through funding from the new National Foundation for Functional Brain Imaging, http://www.nffbi.org/, which is supporting three centers (Minnesota, Mass Gen, and Albuquerque) and whose headquarters are located on the University of New Mexico campus. The lab enjoys a close relationship with the Albuquerque High Performance Computing Center, http://www.ahpcc.unm.edu/, through which we have access to superlative computational resources, and with the nearby Santa Fe Institute, http://www.santafe.edu/. We also have ties to the nearby VA Hospital, which hosts a 122-channel full-head neuromagnetometer. Applications on the Computer Science/Algorithms end of the spectrum should be sent to Barak A. Pearlmutter Computer Science Dept, FEC 313 University of New Mexico Albuquerque, NM 87131 bap at cs.unm.edu and those on the Neuroscience/Experimentation end of the spectrum to Akaysha C. Tang Department of Psychology, Logan Hall University of New Mexico Albuquerque, NM 87131 akaysha at kongzi.unm.edu Applications should include a full cv, names and contact information of three references, and a statement of research interests. The preferred format is emailed plain ascii, but emailed postscript or physical mail are also acceptable. Top candidates will be interviewed on-site. The University of New Mexico main campus is on historic Route 66 in Albuquerque, is an official botanical garden, and is seven minutes by car from a major airport and thirty from a ski resort. The state of New Mexico has the highest per-capita number of PhD's in the country, and offers outstanding outdoor recreational opportunities. For local information see links below http://kongzi.unm.edu/~akaysha/ or http://albuquerque.areaguides.net/. From harnad at coglit.ecs.soton.ac.uk Sun Oct 17 07:56:23 1999 From: harnad at coglit.ecs.soton.ac.uk (Stevan Harnad) Date: Sun, 17 Oct 1999 12:56:23 +0100 (BST) Subject: SELF-CONSCIOUSNESS/Bermudez: PSYC Call for Book Reviewers (844 l.) Message-ID: PSYCOLOQUY CALL FOR BOOK REVIEWERS Below is the Precis of "The Paradox of Self-Consciousness" by Jose Luis Bermudez (799 lines). This book has been selected for multiple review in PSYCOLOQUY. If you wish to submit a formal book review please write to psyc at pucc.princeton.edu indicating what expertise you would bring to bear on reviewing the book if you were selected to review it. (If you have never reviewed for PSYCOLOQUY or Behavioral & Brain Sciences before, it would be helpful if you could also append a copy of your CV to your inquiry.) If you are selected as one of the reviewers and do not have a copy of the book, you will be sent a copy of the book directly by the publisher (please let us know if you have a copy already). Reviews may also be submitted without invitation, but all reviews will be refereed. The author will reply to all accepted reviews. Full Psycoloquy book review instructions at: http://www.princeton.edu/~harnad/psyc.html http://www.cogsci.soton.ac.uk/psycoloquy/ Relevant excerpts: Psycoloquy reviews are of the book not the Precis. Length should be about 200 lines [c. 1800 words], with a short abstract (about 50 words), an indexable title, and reviewer's full name and institutional address, email and Home Page URL. All cited references that are electronically accessible should also have URLs indicated. AUTHOR'S RATIONALE FOR SOLICITING MULTIPLE BOOK REVIEW: The book offers a novel approach to the study of self-consciousness, integrating philosophical argument with detailed study of empirical work from a range of disciplines. It provides a framework for linking together distinct areas of cognitive science which are rarely discussed together and discusses some fundamental problems in the foundations of psychology (such as the nature of concepts and the possibility of thought without language). I am continuing to work on some of the central themes of the book and would greatly benefit from feedback from the biobehavioral and cognitive science community. ----------------------------------------------------------------------- psycoloquy.99.10.035.self-consciousness.1.bermudez Sun Oct 17 1999 ISSN 1055-0143 (47 paragraphs, 30 references, 799 lines) PSYCOLOQUY is sponsored by the American Psychological Association (APA) Copyright 1999 Jose Luis Bermudez THE PARADOX OF SELF-CONSCIOUSNESS (REPRESENTATION AND MIND) Precis of Bermudez on Self-Consciousness [MIT/Bradford, 1998 xiv, 236 pp. ISBN: 0-262-02441-1] Jose Luis Bermudez Department of Philosophy University of Stirling Stirling FK9 4LA Scotland CREA Ecole Polytechnique 1 Rue Descartes 75005 Paris France jose.bermudez at stir.ac.uk ABSTRACT: This book addresses two fundamental questions in the philosophy and psychology of self-consciousness: (1) Can we provide a noncircular account of full-fledged self-conscious thought and language in terms of more fundamental capacities? (2) Can we explain how full-fledged self-conscious thought and language can arise in the normal course of human development? I argue that a paradox (the paradox of self-consciousness) arises from the apparent strict interdependence between self-conscious thought and linguistic self-reference. Responding to the paradox, I draw on recent work in empirical psychology and philosophy to cut the tie between self-conscious thought and linguistic self-reference. The book studies primitive forms of nonconceptual self-consciousness manifested in visual perception, somatic proprioception, spatial reasoning and interpersonal psychological interactions. KEYWORDS: cognitive maps; concepts; content; ecological self; navigation; proprioception; self-consciousness; self-reference; visual perception; I. INTRODUCTION 1. Philosophy and the neurosciences have an uneasy relationship. Fruitful engagement is rare in either direction. This is partly the inevitable result of the division of academic labours. But there is also a deeper reason. The dominant methodological conception governing work in the cognitive sciences involves a distinction of levels of explanation. Marr's theory of vision has often been held up as a model which the cognitive sciences in general ought to follow mainly, of course, because it is one of the very few worked out and satisfying theoretical treatments of a cognitive capacity that cognitive science has so far produced. As is well-known, Marr's approach to the study of the visual system is top-down (Marr 1982). He starts with an abstract specification of the functional tasks that the visual system has to perform, hypothesises a series of algorithms that could compute these functional tasks and then speculates about the implementation of those algorithms at the neural level. Each of the levels of explanation at which the theory operates is relatively autonomous, although of course the computational level models the realisation of the functions identified at the functional level and the implementational level explains how the functions identified at the computational level are realized. The resulting theory is, of course, a dazzling achievement. But there are hidden implications in taking it as a general paradigm for cognitive science. Taking it as a paradigm makes it natural to think, for example, that the place of philosophy is at the functional level and, correspondingly, that the place of cognitive neuroscience is at the implementational level. The result, of course, is that the two disciplines are effectively insulated from each other by the intervening computational level of explanation. 2. There is an obvious problem, however, with generalizing Marr's approach. The problem is that the distinction of levels of explanation really makes sense only where one can identify a clear functional task or set of tasks that need to be carried out. But it is not clear that this can be done outside the restricted domain of encapsulated modules such as the early visual system, the language-parsing system or the face recognition system. Fodor, the most articulate defender of this methodological approach, has clearly appreciated this, and drawn the drastic conclusion that cognitive science cannot hope to shed any light on the so-called central processes of cognition. A more sensible lesson to draw, I think, is that outside this restricted domain a more interactive conception of the relation between the levels of explanation is appropriate. There must be constraints on theorizing at the functional and computational levels. On the top-down approach these constraints emerge from clearly defined functional tasks. But where there are no such functional tasks explanation cannot be purely top-down. There must be constraints and programmatic suggestions moving in both directions. 3. The difficulty in putting this programme into practice is identifying the points of contact between neuroscientific concerns and, for example, philosophical concerns. In this prcis of my book The Paradox of Self-Consciousness (Bermdez 1998) I identify some of the key areas where neuroscientific and philosophical issues intersect in the study of self-consciousness, a form of cognition about as far as it is possible to get from the encapsulated modules where top-down analyses can be so profitably applied. II. THE PARADOX OF SELF-CONSCIOUSNESS 4. In thinking about self-consciousness we need to start with the phenomenon of first-person thought. Most, if not all, of the higher forms of self-consciousness presuppose our capacity to think about ourselves. Consider, for example, self-knowledge, the capacity for moral self- evaluation and ability to construct a narrative of our past. Although much of what we think when we think about ourselves involves concepts and descriptions also available to us in our thoughts about other people and other objects, our thoughts about ourselves also involve an ability that we cannot put to work in thinking about other people and things - namely, the ability to apply those concepts and descriptions uniquely to ourselves. I shall follow convention in referring to this as the capacity to entertain 'I'-thoughts. 5. 'I'-thoughts of course involve self-reference, but it is self-reference of a distinctive kind. Consider the following two ways in which I might entertain thoughts that refer to myself: (1) JLB thinks: JLB is about to be attacked by a poisonous spider (2)JLB thinks: I am about to be attacked by a poisonous spider It is clear that these are very different thoughts, even though they are both thoughts about the same person, namely me. Even if I am suffering a temporary attack of amnesia that has led me to forget my own name I can think the first thought with equanimity. Not so the second. 6. This property of I-thoughts is sometimes described as their immunity to error through misidentification, where this means (roughly) that one cannot think an 'I'-thought without knowing that it is in fact about oneself (Shoemaker 1968, Evans 1982). This feature of 'I'-thoughts is closely tied to the well-known linguistic property of the first-person pronoun, namely, that the first-person pronoun I always refers to the person uttering it. 7. Putting these two properties together suggests the following deflationary account of self- consciousness: (A) Once we have an account of what it is to be capable of thinking 'I'-thoughts we will have explained everything that is distinctive about self-consciousness. (B) Once we have an account of what it is to be capable of thinking thoughts that are immune to error through misidentification we will have explained everything that is distinctive about the capacity to think 'I'-thoughts. (C) Once we have explained what it is to master the semantics of the first person pronoun (e.g. via mastery of some version of the token-reflexive rule that a given utterance of I always refers to the person uttering it), we will have explained everything that is distinctive about the capacity to think thoughts that are immune to error through misidentification. 8. The problem with the deflationary view is that first-person self-reference is itself dependent upon 'I'-thoughts in a way that creates two forms of vicious circularity which collectively I term the paradox of self-consciousness The first type of circularity (explanatory circularity), arises because the capacity for self-conscious thought must be presupposed in any satisfactory account of mastery of the first person pronoun. I cannot refer to myself as the producer of a given token of 'I' without, for example, knowing that I intend to refer to myself - which is itself a self-conscious thought of the type that we are trying to explain. The second type of circularity (capacity circularity) arises because this interdependence rules out the possibility of explaining how the capacity either for self-conscious thought or for linguistic mastery of the first person pronoun arises in the normal course of human development. It does not seem possible to meet the following constraint: The Acquisition Constraint If a given psychological capacity is psychologically real then there must be an explanation of how it is possible for an individual in the normal course of human development to acquire that capacity. Neither self-conscious thought nor linguistic mastery of the first person pronoun is innate, and yet each presupposes the other in a way that seems to imply that neither can be acquired unless the other capacity is already in place. III. ESCAPING THE PARADOX OF SELF-CONSCIOUSNESS 9. The strategy that I employ in the book to escape the paradox of self-consciousness involves making a clear distinction between (a) those forms of full-fledged self-consciousness which presuppose mastery of the first person concept and linguistic mastery of the first person pronoun, and (b) those forms of primitive or nonconceptual self-consciousness which do not require any such linguistic or conceptual mastery. It is these nonconceptual forms of self-consciousness that allow us to escape both the types of circularity Ive just identified. I identified such nonconceptual forms of self-awareness in four domains: (1) perceptual experience (2) somatic proprioception (bodily self-awareness) (3) self-world dualism in spatial reasoning (4) psychological interaction 10. The basic result is that the domain of self-consciousness is far wider than it has been held to be by philosophers. Self-consciousness has often been thought to be the highest form of human cognition, and many philosophers, famous and not so famous, have correspondingly thought that a philosophical account of self-consciousness would be the Archimedean point for a satisfactory account of human thought. But the premise is flawed. Self-consciousness is something we share with prelinguistic infants and with many members of the animal kingdom. The highly conceptual forms of self-consciousness emerge from a rich foundation of nonconceptual forms of self- awareness. As I will try to bring out, recognising this builds a bridge between philosophical interests and neuroscientific ones. IV. THE SELF OF ECOLOGICAL OPTICS 11. One of J. J. Gibsons great insights in the study of visual perception was that the very structure of visual perception contains propriospecific information about the self, as well as exterospecific information about the distal environment (Gibson 1979). Visual perception involves self-perception at the same time as it involves perception of the world. This is the most primitive form of nonconceptual self-awareness, the foundation on which all other forms of self-awareness are built. 12. Gibson stresses certain peculiarities of the phenomenology of the field of vision. Notable among these is the fact that the field of vision is bounded. Vision reveals only a portion of the world to the perceiver at any given time (roughly half in the human case, due to the frontal position of the eyes). The boundedness of the field of vision is part of what is seen, and the field of vision is bounded in a way quite unlike the way in which spaces are bounded within the field of vision. The self appears in perception as the boundary of the visual field a moveable boundary that is responsive to the will. 13. The boundedness of the visual field is not the only way in which the self becomes manifest in visual perception. The field of vision contains other objects that hide, or occlude, the environment. These objects are, of course, various parts of the body. The nose is a particularly obvious example, so distinctively present in just about every visual experience. The cheekbones, and perhaps the eyebrows, occupy a slightly less dominant position in the field of vision. And so too, to a still lesser extent, do the bodily extremities, hands, arms, feet and legs. They protrude into the field of vision from below in a way that occludes the environment, and yet which differs from the way in which one non-bodily physical object in the field of vision might occlude another. They are, as Gibson points out, quite peculiar objects. All objects, bodily and non-bodily, can present a range of solid angles in the field of vision (where by a solid angle is meant an angle with its apex at the eye and its base at some perceived object), and the size of those angles will of course vary according to the distance of the object from the point of observation. The further away the object is, the smaller the angle will be. This gives rise to a clear, and phenomenologically very salient, difference between bodily and non-bodily physical objects. The solid angles subtended by occluding body-parts cannot be reduced below a certain minimum. Perceived body-parts are, according to Gibson, 'subjective objects' in the content of visual perception. 14. But these self-specifying structural invariants provide only a fraction of the self-specifying information available in visual perception. 15. The mass of constantly changing visual information generated by the subjects motion poses an immense challenge to the perceptual systems. How can the visual experiences generated by motion be decoded so that subjects perceive that they are moving through the world? Gibsons notion of visual kinesthesis is his answer to this traditional problem. Whereas many theorists have assumed that motion perception can only be explained by the hypothesis of mechanisms which parse cues in the neutral sensations into information about movement and information about static objects, the crucial idea behind visual kinesthesis is that the patterns of flow in the optic array and the relations between the variant and invariant features make available information about the movement of the perceiver, as well as about the environment. 16. As an example of such a visually kinesthetic invariant, consider that the optical flow in any field of vision starts from a centre, that is itself stationary. This stationary centre specifies the point that is being approached, when the perceiver is moving. The aiming point of locomotion is at the vanishing point of optical flow. 17. Striking experiments have brought out the significance of visual kinesthesis. In the so-called moving-room experiments, subjects are placed on the solid floors of rooms whose walls and ceilings can be made to glide over a solid and immoveable floor (Lishman and Lee 1973). If experimental subjects are prevented from seeing their feet and the floor is hidden, then moving the walls backwards and forwards on the sagittal plane creates in the subjects the illusion that they are moving back and forth. This provides strong support for the thesis that the movement of the perceiver can be detected purely visually, since visual specification of movement seems to be all that is available. An even more striking illustration emerges when young children are placed in the moving room, because they actually sway and lose their balance (Lee and Aronson 1975). 18. The theory of ecological optics identifies a third form of self-specifying information existing in the field of vision. This is due to the direct perception of a class of higher-order invariants which Gibson terms affordances. It is in the theory of affordances that we find the most sustained development of the ecological view that the fundamentals of perceptual experience are dictated by the organism's need to navigate and act in its environment. The uncontroversial premise from which the theory of affordances starts is that objects and surfaces in the environment have properties relevant to the abilities of particular animals, in virtue of which they allow different animals to act and react in different ways. 19. According to Gibson, information specifying affordances is available in the structure of light to be picked up by the creature as it moves around the world. The possibilities which the environment affords are not learnt through experience, and nor are they inferred. They are directly perceived as higher-order invariants. And of course, the perception of affordances is a form of self-perception - or, at least, a way in which self-specifying information is perceived. The whole notion of an affordance is that of environmental information about ones own possibilities for action and reaction. 20. Recognising the existence of the ecological self, as it has come to be known (Neisser 1988), is the first step in resolving the paradox of self-consciousness. It removes the need to explain how infants can bootstrap themselves into the first-person perspective. The evidence is overwhelming that nonconceptual first person contents are available more or less from the beginning of life. Illustrations are to be found in: (1) neonatal distress crying (Martin and Clark 1982) (2) neonatal imitation (Meltzoff and Moore 1977) (3) infant reaching behaviour (Field 1976, Von Hofsten 1982) (4) visual kinesthesis (Lee and Aronson 1984, Butterworth and Hicks 1977, Pope 1984) Let me turn now to some more bottom-level concerns. 21. When we move to considering the neural underpinnings of this form of self-specifying information in visual perception we move into an area that has been fairly closely studied by neuroscientists and experimental psychologists. Particularly relevant here is the proposal, currently under much discussion, that there are two distinct cortical pathways in the human visual system, each carrying distinct types of information (Ungerleider and Mishkin 1982, Goodale and Milner 1992). The distinction between the information carried by the dorsal (infero-temporal) and the ventral (occipito-temporal) pathways respectively has been conceptualized in different ways. Mishkin and Ungerleider see it as a distinction between information about the spatial relations in which an object might stand to the perceiver and information that allows the recognition of objects. Goodale and Milner, in contrast, take the distinction to be between recognitional information about the intrinsic properties of objects (eg their colour, shape and so on) and visuo-motor information about the extrinsic properties of objects (eg their spatial position, orientation, height and so forth). 22. It has been suggested that the action-based self-specifying information that Gibson discusses at the phenomenological level in terms of affordances and invariants in optical flow seems to be carried in the ventral stream (McCarthy 1993). It is far from clear to me, however, that Gibsons insights into the blend of propriospecific and exterospecific information in visual perception fits at all neatly into the proposed distinction of pathways, whether as construed by Mishkin and Ungerleider or by Goodale and Milner at least, if we assume that those processing distinctions are supposed to mark a distinction at the level of conscious phenomenology. The basic concept of an affordance seems to straddle the distinction between where and what, or between recognition and pragmatic. Interestingly, this scepticism about the phenomenological significance of the two cortical pathways is supported by recent work which suggests that the two visual pathways actually collaborate in the control of action (Jeannerod 1997). V. SOMATIC PROPRIOCEPTION AND THE BODILY SELF 23. Gibson's insights into the structure of visual perception were partly vitiated by his insistence on downplaying the importance of somatically derived information about the self. Visual kinesthesis and the perceptual invariants stressed by Gibson are adequate for distinguishing self-movement from movement of the environment, but they are unable to distinguish passive self-movement from active self-movement. They can inform the subject of his movement relative to the environment, but (crudely speaking) they do not tell him whether or not he is moving under his own steam. A different form of self-awareness is required at this point the bodily self-awareness of proprioception. 24. One particularly vivid illustration of the importance of these forms of proprioceptive information comes from the documented cases of complete deafferentation patients who have effectively lost all bodily sensation, either from below the neck in the case of Jonathan Coles patient IW or from below the jaw in Jacques Paillards patient GL (Cole and Paillard 1995). Although IW, unlike GL, can walk, everything he does has to be performed under visual control. Without visual feedback he is incapable of orienting himself and acting. So much so that he sleeps with the light on - if he woke up in the dark he would have no idea where his body was and would never be able to find the light switch. It is interesting, furthermore, to watch a video of him walking. His head is bent forward and pointing downwards so that he can keep his legs and feet in sight constantly. 25. There is a popular sense of self-conscious on which IW seems to be more self-conscious than we are, for the simple reason that everything he does requires his full attention. But this is not the sense of self-consciousness in which I am interested. What is striking about deafferented subjects is how the subjective sense of the body as a bounded spatial entity responsive to the will collapses in the absence of somatic proprioception and can only be partially reestablished with great artificiality and great difficulty. IW and GL are self-conscious in the popular sense precisely because they fail to be self-conscious in a more primitive and fundamental sense. 26. What is this more primitive and fundamental form of self-consciousness that we derive from somatic proprioception? It seems to me to have a tripartite structure. In exploring this it will be useful to start with a list of the principal types of proprioceptive information and their physiological sources. The following is taken from the general introduction to Bermdez, Marcel and Eilan 1995: Information about pressure, temperature and friction from receptors on the skin and beneath its surface. Information about the state of joints from receptors in the joints, some sensitive to static position, some to dynamic information. Information about balance and posture from the vestibular system in the inner ear; the head/trunk dispositional system; and information from pressure on any parts of the body that might be in contact with a gravity-resisting surface. Information about bodily disposition and volume obtained from skin-stretch. Information about nutritional and other homeostatic states from receptors in the internal organs. Information about muscular fatigue from receptors in the muscles. Information about general fatigue from cerebral systems sensitive to blood composition. Information about bodily disturbances derived from nociceptors. 27. At the simplest level, somatic proprioception is a form of self-consciousness simply in virtue of providing information about the embodied self. This is not particularly interesting, although it is worth noting that proprioception gives information about the embodied self that is immune to error through misidentification in the sense discussed earlier. It cannot be the case that one receives proprioceptive information without being aware that the information concerns ones own body. 28. More importantly, somatic proprioceptive information provides a way, perhaps the most primitive way, of registering the boundary between self and non-self. To appreciate this we need to note that there is an important variation among these somatic information systems vary along several dimensions. Some provide information solely about the body (eg. the systems providing information about general fatigue and nutrition). The vestibular system, in contrast, is concerned with bodily balance and hence with the relation between the body and the environment. Other systems can be deployed to yield information either about the body or about the environment. Receptors in the hand sensitive to skin stretch, for example, can provide information about the hand's shape and disposition at a time, or about the shape of small objects. Similarly, receptors in joints and muscles can yield information about how the relevant limbs are distributed in space, or, through haptic exploration, about the contours and shape of large objects. 29. These latter information systems, underpinning the sense of touch, yield a direct sense of the limits of the body and hence of the limits of the self. This is one step further in the development of what might be termed self-world dualism than comes with the self-specifying information in visual perception. The self of visual perception, the ecological self, is schematic and geometrical. Its properties are purely spatial, defined by patterns in the optical flow. It is only in virtue of the sense of touch that the body is experienced as a solid and bounded entity in the world. 30. It is known that a somatotopic map of the surface of the body exists in the somatosensory cortex, and it is natural to think that this plays a key role in subserving the registration of the boundary between self and non-self. Some confirmation confirmation for this will be found in the fascinating work that has been done by V. S. Ramachandran (1994) on somatosensory remapping to explain the well-documented phenomenon of referred sensations in amputees experiencing phantom limbs. The felt boundaries of the body can change as the area in the Penfield homunculus that formerly received input from the amputated limb is invaded by sensory input from nearby areas. 31. The final feature of proprioceptive self-awareness extends this sense of the body as an object. Through feedback from kinesthesia, joint-position sense and the vestibular system we become aware of the body as an object responsive to the will. Proprioception gives us a sense, not just of the embodied self as spatially extended and bounded, but also as a potentiality for action. 32. In this context it might be helpful to point to the role of proprioceptively derived information in the construction of the cross-modal egocentric space within which action takes place. It is well- known that lesions to the posterior parietal cortex produce spatial deficits in primates, human and non-human, and the inference frequently drawn is that the posterior parietal cortex is the brain area where the representation of space is computed. Recent neurophysiological work based on recordings from single neurons has suggested that the distinctive contribution of the posterior parietal cortex is the integration of information from various modalities to generate coordinate systems. Information about visual stimuli is initially transmitted in retinal coordinates. Calibrating this with information about eye position yields head-centred coordinates and further calibration with proprioceptively-derived information yields a body-centred frame of reference. The distal targets of reaching movements are encoded on this modality-free frame of reference, as are motor commands. VI. POINTS OF VIEW 33. The nonconceptual first person contents implicated in somatic proprioception and the pick-up of self-specifying information in visual perception provide very primitive forms of nonconceptual self-consciousness, albeit ones that can plausibly be viewed as in place from birth or shortly afterwards. A solution to the paradox of self-consciousness, however, requires showing how we can get from these primitive forms of self-consciousness to the fully-fledged self-consciousness that comes with linguistic mastery of the first person pronoun. This progression will have to be both logical (in a way that will solve the problem of explanatory circularity) and ontogenetic (in a way that will solve the problem of capacity circularity). Clearly, this requires that there be forms of self- consciousness which, while still counting as nonconceptual, are nonetheless more developed than those yielded by somatic proprioception and the structure of exteroceptive perception and, moreover, that it be comprehensible how these more developed forms of nonconceptual self- consciousness should have 'emerged' out of basic nonconceptual self-consciousness. 34. The dimension along which forms of self-consciousness must be compared is the richness of the conception of the self which they provide. Nonetheless, a crucial element in any form of self- consciousness is the way in which it makes possible for the self-conscious subject to distinguish between self and environment what many developmental psychologists term self-world dualism. In this sense self-consciousness is essentially a contrastive notion. One implication of this is that a proper understanding of the richness of the conception of the self which a given form of self- consciousness provides requires taking into account the richness of the conception of the environment with which it is contrasted. In the case both of somatic proprioception and of the pick- up of self-specifying information in exteroceptive perception, there is a relatively impoverished conception of the self associated with a comparably impoverished conception of the environment. One prominent limitation is that both are synchronic rather than diachronic. The distinction between self and environment that they offer is a distinction that is effective at a time but not over time. The contrast between propriospecific and exterospecific invariants in visual perception, for example, provides a way in which a creature can distinguish between itself and the world at any given moment, but this is not the same as a conception of oneself as an enduring thing distinguishable over time from an environment which also endures over time. 35. To capture this diachronic form of self-world dualism I introduced the notion of a nonconceptual point of view. Having a nonconceptual point of view on the world involves taking a particular route through the environment in such a way that one's perception of the world is informed by an awareness that one is taking such a route. This diachronic awareness that one is taking a particular route through the environment turned out to involve two principal components a non-solipsistic component and a spatial awareness component. 36. The non-solipsistic component is a subject's capacity to draw a distinction between his experiences and what those experiences are experiences of, and hence his ability to grasp that an object exists at times other than those at which it is experienced. This requires the exercise of recognitional abilities involving conscious memory and can be most primitively manifested in the feature-based recognition of places. This is the beginning of an understanding of the world as an articulated, structured entity. 37. The spatial awareness component of a nonconceptual point of view can be glossed in terms of possession of an integrated representation of the environment over time an understanding not just of how the articulated components of the external world fit together spatial, but also of the perceivers own spatial location in the world as a moving perceiver and agent. 38. That a creature possesses such an integrated representation of the environment is manifested in three central cognitive/navigational capacities: The capacity to think about different routes to the same place The capacity to keep track of changes in spatial relations between objects caused by its own movements relative to those objects The capacity to think about places independently of the objects or features located at those places. Powerful evidence from both ethology and developmental psychology indicates that these central cognitive/navigational capacities are present in both nonlinguistic and prelinguistic creatures. 39. This conception of a nonconceptual point of view provides a counterbalance to some important recent work on animal representations of space and their neurophysiological coding. Chapters 5 and 6 of Gallistels The Organization of Learning defend the thesis that all animals from insects upwards deploy cognitive maps with the same formal characteristics in navigating around the environment. Gallistel argues that the cognitive maps that control movement in animals all preserve the same set of geometric relations within a system of earth-centred (geocentric) coordinates. These relations are metric relations. The distinctive feature of a metric geometry is that it preserves all the geometric relations between the points in the coordinate system. Gallistel's thesis is that, although the cognitive maps of lower animals have far fewer places on them, they record the same geometrical relations between those points as humans and other higher animals. Moreover, he offers a uniform acount of how such metric cognitive maps are constructed in the animal kingdom. Dead reckoning (the process of keeping track of changes in velocity over time) yields an earth-centred representation of vantage points and angles of view which combines with current perceptual experience of the environment to yield an earth-centred cognitive map. 40. Without, of course, wishing to challenge Gallistels central thesis that all animal cognitive maps from insects up preserve geometric relations, it nonetheless seems wrong to draw the conclusion that all animals represent space in the same way. Just as important as how animals represent spatial relations between objects is how they represent their own position within the object-space thus defined. And it is here, in what we should think of as not just their awareness of space but also their awareness of themselves as spatially located entities, that we see the major variations and the scale of gradations that the theorists whom Gallistel is criticising have previously located at the level of the cognitive map. VII. PSYCHOLOGICAL SELF-AWARENESS 41. Possession of a nonconceptual point of view manifests an awareness of the self as a spatial element moving within, acting upon and being acted upon by the spatial environment. This is far richer than anything available through either somatic proprioception or the self-specifying information available in exteroceptive perception. Nonetheless, like these very primitive forms of self-consciousness, a nonconceptual point of view is largely awareness of the material self as a bearer of physical properties. This limitation raises the question of whether there can be a similarly nonconceptual awareness of the material self as a bearer of psychological properties. 42. There appear to be three central psychological properties defining the core of the concept of a psychological subject the property of being a perceiver, the property of being an agent, and the property of being a bearer of reactive attitudes. Research on the social cognition of infants shows that there are compelling grounds for attributing to prelinguistic infants in the final quarter of the first year awareness of themselves as bearers of all three of these properties. 43. Psychological self-awareness as a perceiver is manifested in the phenomenon of joint selective visual attention, where infants (a) attend to objects as a function of where they perceive the attention of others to be directed (Scaife and Bruner 1975, Bruner 1975), and (b) direct another individuals gaze to an object in which they are interested (Leung and Reinhold 1981, Stern 1985). In (b), for example, the infant tries to make the mother recognise that he, as a perceiver, is looking at a particular object, with the eventual aim that her recognition that this is what he is trying to do will cause the mother to look in the same direction. 44. Psychological self-awareness as an agent is manifested in the collaborative activities that infants engage in with their care-givers (coordinated joint engagement). Longitudinal studies (e.g. Trevarthen and Hubley 1978) show infants not just taking pleasure in their own agency (in the way that many infants show pleasure in the simple ability to bring about changes in the world, like moving a mobile), but also taking pleasure in successfully carrying out an intention - a form of pleasure possible only for creatures aware of themselves as agents. When, as it frequently is, the intention successfully carried out is a joint intention, the pleasure shared with the other participants reflects an awareness that they too are agents. 45. Psychological self-awareness as a bearer of reactive attitudes is apparent in what developmental psychologists call social referencing (Klinnert et al. 1983). This occurs when infants regulate their own behaviour by investigating and being guided by the emotional reactions of others to a particular situation. The infants willingness to tailor his own emotional reactions to those of his mother presuppose an awareness that both he and she are bearers of reactive attitudes. VIII. CONCLUSION 46. The four types of primitive or nonconceptual self-awareness provide the materials for resolving the paradox of self-consciousness. On the one hand, the problem of capacity circularity can be blunted by showing how it is conceivable that the capacity for full-fledged, conceptual self- consciousness could emerge from the basis of the primitive forms of self-consciousness discussed. On the other, the problem of explanatory circularity can be solved by giving an account of what it is to have mastery of the first-person pronoun that shows how the first-person thoughts involved can be understood at the nonconceptual level. 47. Instead of going into the details of how either of these goals can be achieved, I would like to return to the methodological reflections with which I began. I sketched out what I take to be a dominant approach to the methodology of cognitive science the top-down approach that clearly distinguishes the functional, computational and implementational levels of explanation. As I suggested, this approach really seems applicable only where there are clearly defined identifiable, functional tasks, and consequently is only going to work for peripheral rather than central cognitive processes. The corollary, as Fodor has clearly seen, is that we can expect little illumination of central processes from the cognitive sciences. What Ive tried to sketch out is an alternative approach, one where the distinction of levels of explanation does not correspond to a division of explanatory labour. Ive explored how attending to a particular philosophical puzzle about self- consciousness, perhaps the paradigm central cognitive process, brings out the importance of forms of self-consciousness that look as if they can only be understood by a more interactive collaboration between disciplines whose spheres of competence are so clearly separated on the conventional view. REFERENCES: Bermudez, J. L. 1998. The Paradox of Self-Consciousness. Cambridge MA. MIT Press. Bermudez, J. L., Marcel, A. J. and Eilan, N. (Eds.) 1995. The Body and the Self. Cambridge MA. MIT Press. Bruner, J. S. 1975. 'The ontogenesis of speech acts' in Journal of Child Language 2, 1-19. Butterworth, G. E., and Hicks, L. 1977. 'Visual proprioception and postural stability in infancy: A Developmental Study' in Perception 6, 255-262. Cole, J. and Paillard, J. 1995. 'Living without touch and peripheral information about body position and movement: Studies with deafferented subjects' in Bermudez, Marcel and Eilan (Eds.) 1995. Eilan, N., McCarthy, R. and Brewer, M. W. (Eds.).1993. Spatial Representation: Problems in Philosophy and Psychology. Oxford. Basil Blackwell. Evans, G. 1982. The Varieties of Reference. Oxford. Clarendon Press. Field, J. 'Relation of young infants' reaching behaviour to stimulus distance and solidity' in Developmental Psychology 12, 444-448. Fodor, J. 1983. The Modularity of Mind. Cambridge MA. MIT Press. Gallistel, C. R. 1990. The Organization of Learning. Cambridge MA. MIT Press. Gibson, J. J. 1979. The Ecological Approach to Visual Perception. Boston. Houghton Mifflin. Goodale, M. A. and Milner, A. D. 1992. 'Separate visual pathways for perception and action'. Trends in Neuroscience 15, 20-25. Jeannerod, M. 1997. The Cognitive Neuroscience of Action. Oxford. Basil Blackwell. Klinnert, M. D., Campos, J. J., Sorce, J. F. Emde, R. N. Svejda, M. 1983. 'Emotions as behaviour regulators: Social referencing in infancy' in Plutchik and Kellerman 1983. Lee, D. N., and Aronson, E. 1974. 'Visual proprioceptive control of standing in human infants' in Perception and Psychophysics 15: 529-532. Leung, E. and Rheinhold, H. 1981. 'Development of pointing as a social gesture' in Developmental Psychology 17, 215-220. Lishman, J. R., and Lee, D. N. 1973. 'The autonomy of visual kinaesthetics' in Perception 2: 287-94. Mccarthy, R. A. 1993. 'Assembling Routines and Addressing Representations: An Alternative Conceptualization of 'What' and 'Where' in the Human Brain' in Eilan, Mccarthy and Brewer (Eds.) 1993. Marr, D. 1982. Vision. San Fransisco. W. H. Freeman. Martin, G. B. and Clark, R. D. 1982. Distress Crying in Neonates: Species and Peer Specificity in Developmental Psychology 18, 3-9. Meltzoff, A. N. and Moore, M. K. 1977. 'Imitation of facial and manual gestures by human neomnates' in Science 198, 75-78. Neisser, U. 1988. 'Five Kinds of Self-Knowledge' in Philosophical Psychology, 1, 35-59. Pope, M. J. 1984. Visual Proprioception in Infant Postural Development. PhD Thesis. University of Southampton. Ramachandran, V. S. 1994. Phantom Limbs, Neglect Syndromes, Repressed Memories, and Freudian Psychology in International Review of Neurobiology 37, 291-333. Scaife, M. and Bruner, J. S. 1975. 'The capacity for joint visual attention in the infant' in Nature 253, 265-266. Shoemaker, S. 1968. 'Self-reference and self-awareness' in The Journal of Philosophy 65, 555-567. Stern, D. 1985. The Interpersonal World of the Infant. New York. Basic Books. Trevarthen, C. and Hubley, P. 1978. 'Secondary Intersubjectivity: Confidence, Confiding and Acts of Meaning in the First Year' in LOCK (Ed.) 1978. Ungerleider, M. and Mishkin, L. 1982. Two cortical visual systems. In D. J. Ingle, M. A. Goodale and R. J. W. Mansfield (Eds.), Analysis of Visual Behaviour. Cambridge MA. MIT Press. Von Hofsten, C. 1982. 'Foundations for Perceptual Development' in Advances in Infancy Research 2, 241-261. From harnad at coglit.ecs.soton.ac.uk Sun Oct 17 09:12:56 1999 From: harnad at coglit.ecs.soton.ac.uk (Stevan Harnad) Date: Sun, 17 Oct 1999 14:12:56 +0100 (BST) Subject: PSYC Call for Book Reviewers: Cognitive Mapping/Golledge Message-ID: PSYCOLOQUY CALL FOR BOOK REVIEWERS Below is the Precis of "Wayfinding Behavior: Cognitive mapping and other spatial processes" (685 lines). This book has been selected for multiple review in PSYCOLOQUY. If you wish to submit a formal book review please write to psyc at pucc.princeton.edu indicating what expertise you would bring to bear on reviewing the book if you were selected to review it. (If you have never reviewed for PSYCOLOQUY or Behavioral & Brain Sciences before, it would be helpful if you could also append a copy of your CV to your inquiry.) If you are selected as one of the reviewers and do not have a copy of the book, you will be sent a copy of the book directly by the publisher (please let us know if you have a copy already). Reviews may also be submitted without invitation, but all reviews will be refereed. The author will reply to all accepted reviews. Full Psycoloquy book review instructions at: http://www.princeton.edu/~harnad/psyc.html http://www.cogsci.soton.ac.uk/psycoloquy/ Relevant excerpts: Psycoloquy reviews are of the book not the Precis. Length should be about 200 lines [c. 1800 words], with a short abstract (about 50 words), an indexable title, and reviewer's full name and institutional address, email and Home Page URL. All references that are electronically accessible should also have URLs. ----------------------------------------------------------------------- psyc.99.10.036.cognitive-mapping.1.golledge Sun Oct 17 1999 ISSN 1055-0143 (32 paragraphs, 14 references, 685 lines) PSYCOLOQUY is sponsored by the American Psychological Association (APA) Copyright 1999 Reginald Golledge WAYFINDING BEHAVIOR: COGNITIVE MAPPING AND OTHER SPATIAL PROCESSES. [John Hopkins University Press, 1999 xviii, 428pp, ISBN: 0-8018-5993-X] Precis of Golledge on Cognitive-Mapping Reginald G. Golledge Department of Geography University of California Santa Barbara Santa Barbara CA 93106-4060 U.S.A. golledge at geog.ucsb.edu ABSTRACT: This is an edited volume of essays by psychologists, biologists, cognitive scientists, computer scientists, and geographers on wayfinding by humans and other species. It addresses the extent to which cognitive maps may be universal, and produces evidence that humans, apes, some birds and some small mammals appear to behave as if they have internal representations that guide wayfinding processes in a map-like manner. Evidence also shows that insects, some mammals, and perhaps some birds may not evince such guided behavior, but rely more on spatial updating by dead-reckoning or pilotage. The multiple disciplinary views of wayfinding and navigation by humans and other animals gives the volume a distinctly different content from other available books. KEYWORDS: cognitive map; internal representation; navigation; navigation; path integration; place cells; wayfinding. I. INTRODUCTION 1. The use of maps in forms ranging from dirt drawings to stone carvings, from rice paper scrolls to Automobile Association trip-tiks, from topographic map sheets to disposable tactile strip maps appears to be a cultural universal (Uttal 1997). Maps both record what is known and remembered about an environment and act as wayfinding aids. In the absence of these artifacts, humans and other animals rely on internal representations or stored memories of experienced environments. It is frequently assumed that these stored memories, now commonly referred to as cognitive maps or internal spatial representations, are used to guide travel. Cognitive maps are always there: they cannot be left at home, torn to pieces by fractious children, rendered apart and pieced together incorrectly, so that map reading errors result in a traveler becoming lost. But they do have their problems. 2. The idea that animals also possess internal spatial representations resulted in Tolman's first identification of the term cognitive map. Rather than state that animals, particularly the rats that took short cuts through his mazes stored spatial information as a map, Tolman used the term metaphorically. In other words, he suggested that the animals used in his studies appeared to be able to use spatial information as though the places they remembered were recorded in a maplike manner. For decades, controversy has raged over whether animals do have cognitive maps or if they have other forms of internal spatial representations that allow them to behave as if they were being guided by a map-reading operation. After decades of research in zoology, other biological sciences, and experimental psychology, in particular, various alternatives have been posed to account for successful animal travel behavior. Many have argued that the practice of returning directly to home after a meandering search for food by many nonhuman species indicates that the species did continuous spatial updating, then returned home by a procedure well known to ocean shipping or aircraft pilots the process of dead reckoning. Called path integration, this process enables a traveler to constantly update their current position with respect to an origin without recording details of the path already followed. Because there is no need for a memory trace of the path, route retrace may be difficult or impossible. The need of many foragers who are partly responsible for feeding other members of their species to return home with food, appears to make the short cut (or 'beeline' or 'crow-fly') return trip the more reasonable option. If food is consumed at the spot on which it is found, safety considerations might dictate an immediate shortest-distance return. 3. Flying insects and avian species appear to use landmarks, sun compasses, magnetic compasses, or other celestial guides to help them with migratory and shorter-distance travel. Naturally enough, questions have arisen as to whether the landmarks used are captured as a perspectively viewed retinal image, or whether their configuration or layout is either stored and recalled in sequence as a route is followed, or represented as layouts or configurations similar to a survey (or overview-based) representation of a large-scale and complex space. 4. Despite the existence of these two vigorous research areas, focused on nonhuman and human travel respectively, until recently there have been few deliberate attempts to combine the two literatures. This lack of attention provided the rationale for a small seminar funded by the Borchard Foundation at the Chateau de la Bretesche in July 1996. The purpose of the meeting was to bring together researchers from both the human and nonhuman research domains who had specialized in navigation or wayfinding behavior and who were familiar with the idea of cognitive mapping and the potential role that cognitive maps might play in wayfinding behaviors. As the guests of the foundation's director, William F. Behling and his wife, Betty, at the Chateau de la Bretesche, nine contributors to this volume first presented position statements on the relationships between cognitive maps and wayfinding in humans and other species. To supplement this group's expertise, other scientists were invited to add chapters to the book. 5. The term 'cognitive maps' is used throughout this book to refer to the internal spatial representation of environmental information. Its use varies, from the metaphorical ('as if' the information was stored in maplike format), to a hypothetical construct. 6. The term 'spatial representation' is also used throughout the book. This might be regarded as a shorthand notation for the organization of components of spatial knowledge or other partly investigative processes (e.g., neurophysiological structures and place cells, cell assemblies, phase sequences). The term also can be used metaphorically, involving an 'as if' quality, particularly when referring to purported maplike properties of representations. It has also been used as an intervening variable in which it is interpreted as a 'note' attached to an economical grouping of measured variables in a statement of functional relations between other measured variable. 7. Structurally, this book is divided into four sections, ranging from wayfinding and cognitive mapping in humans operating in different scenarios, to examinations of special human navigation processes (e.g. without sight), to studies of wayfinding by various non-human species, and the neurobiological bases of environmental knowledge acquisition and use. Each section is now summarized in turn. II. SECTION I: HUMAN COGNITIVE MAPS AND WAYFINDING 8. This first section explores the strong theoretical and empirical links between cognitive maps (or the internal representation of environmental information); the cognitive mapping process itself; the internal manipulation of information in the form of spatial choice and decision making, and the directed acts of human wayfinding through simple or complex environments. The evidence is clear and overwhelming that human wayfinding is directed and motivated, and follows sets of procedural rules whose content and structure are the focus of much ongoing research. The consensus is clear: humans acquire, code, store, decode, and use cognitive information as part of their navigation and wayfinding activities. Although over the centuries they have developed numerous ways of supplementing personally stored environmental information (e.g., maps, written descriptions, and various forms of image representations), it appears that humans rely on personal cognitions to make many spatial decisions, and to guide their movement behavior. There is evidence that internal representations and their externalizations (spatial products) do not necessarily match well, and that the existence of fragmented, incomplete, or distorted cognitive maps appears to account for many behaviors that might otherwise be labeled as spatially irrational. 9. The purpose of this part of the book is to examine sets of concepts deemed relevant to both human wayfinding and cognitive mapping. There are two chapters by geographers (Golledge, and Stern and Portugali), and two by psychologists (Allen and Grling). Although some disciplinary perspectives are evident, there is much overlap and common concern. The first two chapters, by Golledge and Allen (respectively), provide overviews and summaries of theories and concepts relevant throughout the entire book. The following chapters by Grling, and Stern and Portugali have a tighter focus: Grling emphasizes the sequential spatial choice processes so important to human wayfinding, and Stern and Portugali emphasize decision making in urban environments, the complex scenarios in which most humans live and interact. All four chapters contain examples of relevant research. 10. In the first chapter, Golledge reviews critical definitions relating to cognitive maps and wayfinding. He provides an overview of the role of cognitive mapping in human wayfinding and describes the processes of acquiring and storing spatial information about large-scale complex environments. Further, he discusses how humans record and represent environmental knowledge. The role played by landmarks and routes in anchoring knowledge and in wayfinding is examined, and the differences between path following and route-based environmental learning are explored. Errors commonly related to encoding, decoding and internally manipulating cognized spatial data are highlighted. Wayfinding by humans in contexts other than with landmark usage is also examined, and an elaboration of errors commonly found in human wayfinding follows. Throughout, ties are made to treatments of similar problems in later chapters that focus on the nonhuman domains of internal spatial representations and wayfinding. 11. In the second chapter, Allen provides insights into the nature of spatial abilities and the role they play in cognitive mapping and wayfinding procedures. He places emphasis on the concept of individual differences in spatial cognition and in behavior. Allen argues that the scientific literature in psychology and geography contains a vast number of studies concerned with spatial abilities and a growing body of research on wayfinding, although little has been done to establish the relevance of the former for the latter. Thus the question of why some individuals are better than others at wayfinding has been difficult to address. Allen suggests that a potentially informative way to think of wayfinding is to differentiate between wayfinding tasks and wayfinding means. Tasks include traveling to a previously known destination, exploration with the purpose of returning home, and traveling to a novel destination. Means include oriented search, following a continuously marked trail, piloting (between landmarks), habitual locomotion, path integration, and reference to a cognitive map. Spatial abilities in the past have been examined from psychometric, information processing, developmental, and neuropsychological perspectives. Allen suggests that broad fami1ies of abilities involved in the identification of manipulable objects, those involved in anticipating the trajectory and speed of moving objects, and those involved in supporting oriented travel within large-scale environments summarize the dominant research themes. He implies there is considerable utility associated with the concept of interactive common resources for cognitive and perceptual-motor tasks. The result of the use of spatial abilities is support-oriented travel, but they also serve as a resource for acquiring additional environmental knowledge. Cognitive maps are considered as knowledge of places and cognitive mapping includes rules for establishing spatial relations among such places. 12. Next, Grling discusses human information processing in sequential spatial choice, which summarizes the essential acts involved in wayfinding. He begins with the premise that human locomotion in space is goal-directed. Spatial orientation and navigation are, therefore, primarily means of monitoring travel plans. Travel plans are developed prior to initiating movement. The chapter focuses on the formation of travel plans and their consequent execution. Such planning entails spatial choices that are multiattribute, sequential, and stated. He summarizes research on how people process information when solving the traveling salesman problem (i.e., finding the shortest distance between an origin and a set of destinations that might be sequentially visited). He details research on how time and priority are traded off against spatial attributes in sequential spatial choices. 13. Stern and Portugali next examine the relationship between environmental cognition and decision making in urban navigation. They define urban navigation as a sequential process of decision making concerning route choice. They claim that traditionally many choice situations are described by a 'black-box approach', which does not specify choice rules but rather deals only with the relationship between input and output variables. In most of these models a cognitive explanatory mechanism of the choice process is missing. Their chapter presents two complementary conceptual frameworks as possible ways to solve this problem. The first is the inter-representational network (IRN), and the second is decision field theory (DFT). It is suggested that both frameworks can explain the dynamics and high variability in the choices of persons navigating in urban environments. 14. In the exploration of human wayfinding and its various components as illustrated in this first section, the importance of individual differences, those between males and females, and variations according to one's spatial abilities are reviewed. III. SECTION II: PERCEPTUAL AND COGNITIVE PROCESSING OF ENVIRONMENTAL INFORMATION 15. In this part, three chapters explore cognitive processes and human navigation in a variety of contexts, including an extensive investigation of path integration by humans covering wayfinding without vision; updating an object's orientation and location during nonvisual navigation; exploring the geometrical constraints and calibration of action-representation couplings, and relating perceptual processes to various navigation requirements. Focusing primarily on aspects of human perception and cognition with respect to wayfinding, these authors explore nontraditional domains to show the versatility of relevant theories and concepts. Although vision is accepted as the most important spatial sense, there is no doubt that blind or vision-impaired humans can become competent independent travelers using simple cognitive processes and aids such as the white cane, guide dog, or a variety of recently developed auditory navigational aids. 16. In the first of these chapters (chapter 5) Loomis, Golledge, Klatzky, and Philbeck discuss the process of human navigation by path integration, a process that until recently was recognized more in the nonhuman domain. They begin by clearly defining two types of processes influential in wayfinding piloting and path integration. The recent literature is replete with misconceptions of the nature of these processes, but little is left in doubt following their clear and comprehensive discussion. Navigation by humans, animals, and machines is accomplished using two distinct methods. Piloting is the determination of current position and orientation using landmark information in conjunction with a map, either external or internal. Path integration is the updating of position and orientation on the basis of velocity and acceleration information about self-movement. The chapter begins with a consideration of a number of models of path integration. Following is a review of the empirical research on human path integration, with a focus on controlled experimental investigations. Such investigations have been carried out using two distinct tasks: return-to-origin after the passively guided traverse of an outbound path, and perceptually directed action, whereby the person sees or hears a target and then, with the target extinguished, attempts to indicate its position by actively locomoting toward it or by pointing in its direction during locomotion that passes by the target. 17. In Chapter 6, Amorim discusses a neurocognitive approach to human navigation. He suggests that human navigation is viewed as a result of the interplay of neurocognitive functions. Spatial updating and frames of reference constitute the two concepts of maximum interest in this work. He provides experimental evidence on the role of reference frames in computing locations in space, as well as on the effect of two processing modes for the updating of an object's location and orientation. Amorim uses an information-processing approach (commonly used in cognitive psychology) in an effort to understand human processes of updating an object's location and orienting it with respect to a bounding frame of reference. To localize a person in the environment as well as localize an object the environment contains, Amorim suggests that the acquisition, coding, and integration of sensory information (both perceptual and representational) are necessary. Building on the model of visuo-spatial cognition proposed by Kosslyn (1991), Amorim offers two studies; one investigates the role of reference frames in computing locations in space, whereas the other compares two processing modes for the updating of an object's location and orientation in space. In interpreting the results of these experiments he evaluates the neurocognitive approach to the study of the pathological causes of topographical disorientation. 18. In chapter 7, Rieser examines action-representation couplings, focusing in particular on the geometrical constraints on such calibrations. He argues that perception and action are coupled, so that motoric actions result in dependable changes in the actor's perspective. For example, during locomotion the structure of an actor's perspective visibly rotates and translates in directions and at rates that fit with the geometry and rate of locomotion. This coupling provides a chance for perceptual-motor learning. While walking with vision, people learn the covariation of optical (and possibly nonoptical) flow and afferent-efferent input associated with the biomechanical activities of walking. This learning, in turn, provides the basis for the coupling of representation and action. Representation is coupled with action in working memory in analogous ways. When acting without vision, people are knowledgeable about the resulting changes in their perspective. So for example, after viewing their surroundings and then walking without vision, people are able to keep up to date on the changing self-to-object distances and directions relative to their remembered surroundings. IV. SECTION III: WAYFINDING AND COGNITIVE MAPS IN NONHUMAN SPECIES 19. In previous sections we focused on humans, in whom cognitive processing is well established, but the tie to navigation and wayfinding is not strongly defined. In this section the authors focus on navigation and wayfinding by nonhuman species, in which the presence of cognitive maps is being strongly debated. In these chapters, biological and ecological scientists examine wayfinding and discuss the possibility that different species have and use cognitive maps. 20. Etienne, Maurer, Georgakopoulos, and Griffin begin (in chapter 8) with a review of the significance of dead reckoning or path integration and landmark use in the representation of space. In many ways this provides a view that complements chapter 5 by Loomis et al., which presents a human navigator's view of the same process. In particular they examine suggestions that dead reckoning (which does not involve learning an environment) seems more dominant in nonhumans, whereas landmark-guided movement may be more dominant in humans. The problem of how different species combine the systems in wayfinding is examined in great detail. Drawing on examples from their group's work with small mammals, Etienne et al. suggest that animals may well have a simple cognitive map that helps their memory for routes and places (such as sources of food or food storage areas). 21. But not all animals may have such cognitive maps. In this chapter, Etienne et al. begin from the viewpoint that spatial representation as defined originally by Tolman (1948) and more recently by O'Keefe and Nadel (1978) refers to a high level of spatial information processing. They use the term cognitive map to imply that a subject organizes the familiar environment as a system of interconnected places and that it applies a set of transformation rules to this system, which may consist of a limited number of complementary operations (such as those hypothesized by Piaget 1937), or that optimize goal-directed movements. Thus whether human or nonhuman, a subject must be able to pilot and perform new route selection before being credited with possessing a cognitive map. 22. The authors define piloting in terms of planning and performing a goal-directed path by deducing an itinerary from the memorized spatial relations between a goal and a traveler's current position, while new route performance implies an ability to select the most economical alternative path (including shortest path and shortcuts) in both familiar and unfamiliar settings. If a cognitive map alone is used, then piloting and path following must take place without either the use of beacons or reference to external landmarks. Etienne et al. argue that the general literature has yet to yield convincing evidence that spatial knowledge reaches this degree of coherence in species other than primates. They suggest using the term spatial representation, or more precisely, the representation of locomotive space, for their work with nonprimate animal species. Thus their chapter directly addresses the question of the universality of cognitive maps by suggesting that whereas spatial representation may be universal, cognitive maps may develop only in a limited number of species. They then point out that the attribution of specific systems of representation to different species poses severe problems. They argue that if one ascribes to an animal or a young child particular forms of spatial representation, inevitably one begins by analyzing subjects' behavior in specific functional contexts to see how observed behaviors fit certain aspects of the environment. 23. The authors make a strong statement that all sedentary species adapt their locomotor behavior to relevant features in the spatial environment in order to reach their goals without getting lost. Thus the observed correspondence between behavior and functionally meaningful aspects of the environment gives insights into what the traveler knows about the environment and thus how the external world is represented or modeled. The authors then examine the process of dead reckoning, with and without the possible use of ancillary landmarks. They report that many theories of navigation emphasize that dead reckoning (path integration) plays a significant role in spatial representation and wayfinding across the entire animal kingdom from insects and other invertebrates to mammals (Gallistel 1990). Then, building on this fascinating introduction, the authors examine the role of dead reckoning in the representation of space in a comparative perspective, including hymenopterans and rodents. They describe how insects and mammals use dead reckoning as current route-based information and how they use landmark-place associations as long-term location-based references. They then consider how the species previously mentioned represents space on the basis of route-based and location-based information, and on the interaction between these two categories of references. 24. In chapter 9, Judd, Dale, and Collett examine the fine structure of view-based navigation in insects. They begin by asserting that insects learn landmarks as two-dimensional views. These views are highly dependent on vantage points, so that even over a relatively short section of a foraging trip, the insect's view of a nearby landmark will change appreciably. Insects simplify the problems of using such retinotopic views for navigation in a number of ways. For example, bees and wasps restrict the range of directions in which they approach a familiar place so that they capture roughly the same sequence of retinal images from visit to visit (i.e.. approach from the same perspective view). They are guided into the vicinity of the goal by aiming at a nearby beacon landmark. Because of changes in image size and shape, a single stored view of the beacon is unlikely to allow the insect to recognize it over the whole range of possible approaches. In addition, the authors claim that wood ants are shown to take several 'snapshots' of a beacon at different distances in the early stages of learning a new environment. Once close to a beacon, the insect relinquishes fixation either to approach another beacon or to approach the goal. This transition is achieved by linking a stereotyped action to a frontally stored view of the beacon. By this means the insect can acquire a standard view of the next beacon or arrive at a point close enough to the goal to allow image matching of the goal itself or the nearest landmark. The goal is then pinpointed by moving so that the image on the retina matches the view of nearby landmarks. The authors go on to suggest that there is surprising similarity in the motor constraints and landmark strategies of real insects and those of simple simulated 'creatures' they have 'evolved' artificially. Again, the parallel between human and non-human species stands out. 25. Moving from ground-based animals and low-flying insects to birds, the internationally acclaimed team of Wolfgang and Roswitha Wiltschko (chapter 10) discuss compass orientation and basic elements in avian orientation and navigation. Birds face orientation tasks in two behavioral contexts: homing and migration. Because of the long distances involved in migration, birds must establish contact to their goal indirectly via an external reference. Three such mechanisms have been described: a magnetic compass based on the field lines of the geomagnetic field and two compass mechanisms based on celestial cues, namely a sun compass and a star compass. To use a compass, birds must first determine the compass course leading to their destination. For homing, experimental evidence indicates that experienced pigeons can derive the home course from site-specific information obtained at the starting point of the return flight. Their ability to do this even at distant, unfamiliar sites has led to the concept of the navigational 'map', which is a directionally oriented representation of the distribution of environmental gradients within the home region. It can be extrapolated beyond the range of direct experience. Birds determine their home course by comparing local values of these gradients with the home values. The 'map' is based on individual experience. 26. During an early phase in life, young pigeons derive their home course from directional information collected during the outward journey. On spontaneous flights, they record prominent landmarks and changes in navigational factors and combine this information with the direction flown to form the navigational map. Once the map is established, it is preferentially used, because it permits the correction of errors. The navigational map is a cognitive map because it allows novel routes; it differs from cognitive maps discussed for other animals by the size of the area covered and by including continuous factors like gradients. In migration, birds must reach a distant region of the world. The course leading to this goal area is constant; the birds possess genetically coded information on their migratory direction. The conversion of this information into an actual compass course requires external references, which are provided by celestial rotation and by the geomagnetic field. Celestial rotation indicates a reference direction away from the celestial pole, whereas the magnetic field defines a specific deviation from this course, resulting in the population-specific migration course. Both types of cue continue to interact during migratory flights. Depending on the nature of the orientation tasks, birds make use of innate information or of individual learning processes. In both strategies, however, external references provided by compass mechanisms are essential components. 27. Thinus-Blanc and Gaunet (chapter 11) discuss the cognitive map as an internal representation of an environment where places and their spatial relationships (such as angles and distances) are charted. This notion has been extensively criticized in the past by Thinus-Blanc because the expression is antinomic and can easily lead to misunderstanding. The authors point out that 'cognitive' refers to dynamic processes and 'map' refers to a static picture of the real world. To this extent, the term cognitive mapping is functionally more correct. Internal spatial representations are said to be useful for orienting in a given environment just as they contribute to the organization of new spatial information as it is accrued. Thus Thinus-Blanc and Gaunet argue that spatial representation may be viewed as maps of the environment but more appropriately should be viewed as cognitive or active information seeking structures. They draw on data from animal and human studies and related theoretical work to support this hypothesis. V. SECTION IV: THE NEURAL AND COMPUTATIONAL BASES OF WAYFINDING AND COGNITIVE MAPS 28. In this part cognitive neuroscientists Nadel (University of Arizona) and Berthoz (Laboratoire de Physiologie de la Perception et de l'Action, College de France) respectively examine the neural bases of wayfinding and cognitive maps, and computer scientist Chown (Bowdoin College) discusses their implications for computation and artificial intelligence-based travel. 29. In chapter 12, Nadel provides an overview of the neural mechanisms of spatial orientation and wayfinding. He suggests that work on the neural bases of wayfinding in mammals has intensified in recent years, building on the discovery of place neurons in the hippocampus. The cognitive map theory of hippocampal functioning, first put forward by O'Keefe and Nadel in 1978, suggests that this brain structure is the core of an extensive neural system subserving the representation and use of information about the spatial environment. Nadel argues that evidence supporting this theory comes primarily from brain lesion and neurophysiological recording studies. The former showed that damage in the hippocampus system invariably impairs the ability of animals and humans to learn about, remember, and navigate through environments, while the latter show that neurons in this system code for location, direction, and distance, thereby providing the elements needed for a mapping system. Current work in this area focuses on which stimuli control the activity of these neural elements, and how the system is used in behavior. He cites the fact that the roles of external and internal sources of information are under active investigation. 30. In the next chapter, Berthoz and his associates examine the neural bias of spatial memory during locomotion. This chapter addresses the question of the mechanisms that underlie the capacity to memorize routes and to use this spatial memory for guiding and steering of locomotion. A review is presented of several paradigms used in their laboratory to study this question. First, some previous studies, which have shown that vestibular information about head rotation and translation can be used by the brain to estimate distances are reviewed. Berthoz et al. claim that such use has been shown by the vestibular memory contingent 'saccade task', both in normal subjects and in neurological patients. Second, some recent experiments that use the task of walking along a triangular path with or without vision are described. During this task, head position and velocity are measured by video-computerized techniques. Two main results have been obtained: (1) They have discovered that the head anticipates the body movement during walks around a corner: this anticipation also exists in darkness, suggesting that the orienting system is driven by an internal representation of the trajectory and that the brain uses a strategy of guiding locomotion by gaze (go where you look) even in darkness. (2) When vestibular-deficient patients perform the task, they seem to control the total distance but not the direction, suggesting a dissociation between the control of distance and direction in this locomotor pointing task. They also describe a second paradigm of circular locomotion, during which subjects were asked to walk around a circular path with or without vision. Here again, the measure of the kinematics parameters of head movement indicates both an anticipation of head direction and a dissociation between the control of distance and direction, and provides clarification of the frequently misinterpreted concepts of course and heading. Finally, they review a number of recent results which may lead to an understanding of the neural basis of both anticipation and the role of vestibular cues in the steering of locomotion. 31. In the final chapter, Chown discusses error tolerance and generalization in cognitive maps. He asserts that human cognitive maps are not precise, complete, nor necessarily accurate. Because navigation is so important in everyday life, it is not easy to understand why humans have evolved an internal representation of space that appears to have such basic flaws. The theme of this chapter is that it is exactly the sketchy nature of human cognitive maps that make them such a powerful tool for navigation. There is growing evidence from artificial intelligence and robotics that in real environments, useful representations cannot be achieved without sacrificing completeness and precision. Further, it can be shown that the sketchy nature of cognitive maps more naturally lends itself to error tolerance and generalization than would be the case with alternative structures. Cognitive maps may be sketchy but the information they do store is usually sufficient for human needs. The relationship between human needs and how cognitive maps encode information is discussed in a proposed model called PLAN. VI. FINAL COMMENT: 32. The more we know about how humans or other species can navigate, wayfind, sense, and record and use spatial information, the more effective will be the building of future guidance systems, and the more natural it will be for humans to understand and control those systems. The question of which of the many cognitive mapping, navigational, or wayfinding procedures and behaviors should be taken as the role model for future systems remains unanswered at this stage. Knowing the advantages and disadvantages, the strengths and the shortcomings, the idiosyncrasies and the universals of spatial knowledge acquisition and storage and wayfinding behavior can only lead to the development of systems that are as endemic as path integration, as powerful as cognitive mapping, and as anchored as landmark usage, and that possess the versatility to handle both view-centered and object-centered modes of recording or experiencing new environments. REFERENCES: Gallistel, C. R. (1990). The organization of learning. Cambridge, MA: MIT Press. Kosslyn, S. M. (1991). 'A cognitive neuroscience of visual cognition: Further developments.' In R. H. Logie & M. Denis (Eds.) Mental Images in Human Cognition (pp.351-381). Amsterdam: Elsevier Science Publishers. O'Keefe, J., & Nadel, L. (1978). The hippocampus as a cognitive map. Oxford: Oxford University Press. Piaget, J. (1937). La construction du reel chez l'enfant. Paris: Delachaux et Niestl, Neuchtel. Tolman, E. C. (1948). 'Cognitive maps in rats and men.' Psychological Review, 55, 189-208. Uttal, D.H. (1997, April) 'Seeing the big picture: Children's mental representation of spatial information acquired from maps.' Paper presented at the 93rd annual meeting of The Association of American Geographers, Ft. Worth, TX. List of author names with chapter titles and page numbers: 1: Human Wayfinding and Cognitive Maps, 5-46. Reginald G. Golledge 2: Spatial Abilities, Cognitive Maps, and Wayfinding: Bases for Individual Differences in Spatial Cognition and Behavior, 46-81. Gary L. Allen 3: Human Information Processing in Sequential Spatial Choice, 81-99. Tommy Grling 4: Environmental Cognition Decision Makaing in Urban Navigation, 99-121. Eliahu Stern Juval Portugali PART II PERCEPTUAL AND COGNITIVE PROCESSING OF ENVIRONMENTAL INFORMATION 5: Human Navigation by Path Integration, 125-152. Jack M. Loomis Roberta L. Klatzky Reginald G. Golledge John W. Philbeck 6: A Neurocognitive Approach to Human Navigation, 152-168. Michel-Ange Amorim 7: Dynamic Spatial Orientationa and the Coupling of Representation and Action, 168-191. John J. Rieser PART III WAYFINDING AND COGNITIVE MAPS IN NONHUMAN SPECIES 8: Dead Reckoning (Path Integration), Landmarks, and Representation of Space in a Comparative Perspective, 197-229. Ariane S. Etienne Roland Maurer Josephine Georgakopoulos Andrea Griffin 9: On the Fine Structure of View-Based Navigation in Insects, 229-259. Simon P. D. Judd Kyran Dale Thomas S. Collett 10: Compass Orientation as a Basic Element in Avian Orientation and Navigation, 259-294. Roswitha Wiltschko, Wolfgang Wiltschko 11: Spatial Processing in Animals and Humans: The Organizing Function of Representations for Information Gathering, 294-309. Catherine Thinus-Blanc Florence Gaunet PART IV THE NEURAL AND COMPUTATIONAL BASES OF WAYFINDING AND COGNITIVE MAPS 12: Neural Mechanisms of Spatial Orientation and Wayfinding: An Overview, 313-328. Lynn Nadel 13: Dissociation between Distance and Direction during Locomotor Navigation, 328-349. Alain Berthz Michel-Ange Amorim 14: Error Tolerance and Generalization in Cognitive Maps: Performance without Precision 349. Eric Chown REFERENCES 371 CONTRIBUTORS 415 INDEX 419 From curt at cassandra.ucr.edu Mon Oct 18 18:45:22 1999 From: curt at cassandra.ucr.edu (Dr. Curt Burgess) Date: Mon, 18 Oct 1999 15:45:22 -0700 Subject: quant & developmental cognition position at Univ Calif, Riverside Message-ID: Just wanted to make sure that any interested parties on the Connectionist email distribution list knew about this position. Curt Burgess ---------------------- The Department of Psychology, University of California, Riverside, invites applications for a tenure-track Assistant Professor position in Developmental Psychology, beginning July 1, 2000. The Ph.D. degree is required at the time of the appointment. Applicants in all areas of developmental psychology are encouraged to apply. Preferred applicants are those with an interest in cognitive development with strong quantitative skills, who can contribute to the graduate program in quantitative psychology. We are seeking a developmental psychologist with a strong program of substantive research that complements that of the current developmental faculty, whose emphases include cognitive and social development, and processes of socialization. Applicants should also be committed to excellence in undergraduate and graduate education. The deadline for review of completed applications begins January 10, 2000 and continues until the position is filled. Interested candidates should send their curriculum vitae, reprints if available, a cover letter describing research and teaching interests, and arrange to have three letters of recommendation sent to: Chair, Developmental Psychology Search Committee Box I-CN Department of Psychology University of California - Riverside Riverside, CA 92521 The Riverside campus of the University of California is growing rapidly and has an excellent psychology department with a strong record of success in research, teaching, and extramural funding. For information on the Department of Psychology, see our web site at: http://www.psych.ucr.edu/. The campus is centrally located in Southern California, about 50 miles east of Los Angeles and less than an hour's drive from the area's mountains, deserts, and beaches. The University of California, Riverside is an equal opportunity employer/affirmative action employer. --------------------- Dr. Curt Burgess, Computational Cognition Lab Chair of Graduate Admissions for Psychology Department of Psychology, University of California 1419 Life Science Building Riverside, CA 92521-0426 URL: http://locutus.ucr.edu/ Internet: curt at cassandra.ucr.edu MaBellNet: (909) 787-2392 FAX: (909) 787-3985 From wolpert at hera.ucl.ac.uk Tue Oct 19 06:48:25 1999 From: wolpert at hera.ucl.ac.uk (daniel wolpert) Date: Tue, 19 Oct 1999 11:48:25 +0100 (GMT) Subject: Positions in Computational Sensorimotor Control Message-ID: <14348.19526.664653.987081@perseus.ion.ucl.ac.uk> ********************************************************************* Nature advertisement October 21st Institute of Neurology, University College London Applications are invited for three posts to work with Dr. Daniel Wolpert under a Wellcome Programme Grant entitled Computational Human Sensorimotor Control. These positions are available for three years in the first instance with a starting date from January 2000, but candidates who wish to start later in 2000 are also encouraged to apply. Further details of all posts are available on http://www.hera.ucl.ac.uk TWO POSTDOCTORAL RESEARCH FELLOWS Two Fellows are required with PhDs relevant to the study of human motor control. One Fellow will work on modelling of the motor system using optimal control and connectionist approaches and should have a background in a computational field (e.g. Computational Neuroscience, Engineering, Physics, Maths). One Fellow will work on psychophysical studies of human motor control and should have a background in an experimental field (e.g. Neuroscience, Psychology). RESEARCH ASSISTANT The Research Assistant will provide software and hardware support for the group. Experience with C/C++ programming essential and experience with NT, Linux, OpenGL, external device interface/control and electronics would be advantageous. Starting salary for all positions is up to ?27,586 pa inclusive, `depending on experience. Applications (2 copies of CV and names of 3 referees) to Miss E Bertram, Assistant Secretary (Personnel), Institute of Neurology, Queen Square, London WC1N 3BG (fax: +44 (0) 171 278 5069) by 1 December 1999. Informal enquiries to Dr. Daniel Wolpert (email: wolpert at hera.ucl.ac.uk; tel: 0171 837 3611 ext 4183; fax: 0171 813 3107). Working toward Equal Opportunity ********************************************************************* From Richard.Woesler at t-online.de Wed Oct 20 10:00:16 1999 From: Richard.Woesler at t-online.de (Richard Woesler) Date: Wed, 20 Oct 1999 16:00:16 +0200 Subject: Thesis Announcement Message-ID: -------------------------------------------------------------------------------- Dear Connectionists, I am pleased to announce the availability of my thesis. Regards, Richard Woesler -------------------------------------------------------------------------------- Title: Neural Networks for Object Segmentation in Amphibians and an algorithmic Classification of Neurons ISBN 3-89722-073-3, Logos-Verlag Berlin, Germany 1999 Advisors: Prof. Helmut Schwegler, head of the Institute of Theoretical Neurophysics at the University of Bremen, and Prof. Gerhard Roth, head of the Institute of Brain Research at the University of Bremen, Germany Language: german Abstract: Neural network models for object segmentation in amphibians, especially in the tongue projecting salamander Hydromantes italicus, constitute the main part of the text. Firstly, we carry out a simple double-dummy experiment in which the animal has to choose one dummy out of two identical, coherently moving prey-dummies. Hydromantes is able to select one of them and to snap at it with high accuracy. The question arises how this object-selection process, the segmentation of the selected object from the rest of the visual scene, and the high accuracy are realized within the amphibian brain. First, we present a simplified model, our Base Model. This model detects moving contrast edges, binds these edges to objects, and selects one object. We derive analytical results concerning the performance level of the network. The results are similar to those known from amphibians. Second, we extend this model to our Amphibian Model where a number of recent biological data is taken into account, especially data about the retina and the tectum opticum which is an important brain region with respect to prey-capture behaviour and which is homologous to the mammalian superior colliculus, and data about the nucleus isthmi which is a relatively small brain region connected reciprocally with the tectum opticum and which is homologous to the mammalian nucleus parabigeminalis. This model is a complete neural network for object segmentation from retinal photoreceptors to the segmentation of the selected object. The model yields possible explanations for many biological findings: e. g., for the width distribution of the receptive fields of neurons of the tectum opticum, or for the range of velocities of objects for which Hydromantes italicus shows prey-capture behaviour. The model contains a spotlight network which enhances the retinotectal transmission in a certain spotlight region to further neurons, which we call the multiplicative coarse coding neurons. The spotlight network should segment the selected object from the rest of the visual scene, and the multiplicative coarse coding neurons should encode its place with high accuracy. We describe in detail how this network can be realized in amphibians. Third, we present the computer model Simulander III with which we simulate the described double-dummy experiment. Indeed the model shows the high object-selection ability and the high accuracy known from Hydromantes. We give an outlook to various model extensions, especially to a model of recognition of the absolute width of the selected object and of recognition of a stepwise prey movement. These abilities are also known from Hydromantes. Additionally, we discuss a model extension which includes a neural network of synchronizing neurons. In the second part of the text, we consider the classification of neurons of the tectum opticum of Hydromantes italicus according to their responses to different prey-dummies. We are particularly interested in the question whether the response types form distinct classes or a continuum. We classify the data with an algorithm that we have developed using known classification methods. The classification result suggests that the data form a continuum with some accumulations. Finally, we discuss the possible biological reasons for a continuum and possible advantages of a continuum. -------------------------------------------------------------------------------- ******************************************** Dipl.-Phys. Dr. Richard Woesler August-Schlueter-Str. 39 48249 Duelmen Germany Phone +49 (0) 2594-991182 Fax +49 (0) 2594-948752 http://home.t-online.de/home/Richard.Woesler ******************************************** From giles at research.nj.nec.com Wed Oct 20 19:29:34 1999 From: giles at research.nj.nec.com (Lee Giles) Date: Wed, 20 Oct 1999 19:29:34 -0400 (EDT) Subject: call for participation - NIPS99 Message-ID: <199910202329.TAA10434@alta.nj.nec.com> CALL FOR PARTICIPATION -- NIPS*99 http://www.cs.cmu.edu/Web/Groups/NIPS Neural Information Processing Systems -- Natural and Synthetic Monday November 29 - Saturday December 4, 1999 Denver, Colorado This is the thirteenth meeting of an interdisciplinary conference which brings together cognitive scientists, computer scientists, engineers, neuroscientists, physicists, and mathematicians interested in all aspects of neural processing and computation. The conference will include invited talks as well as oral and poster presentations of refereed papers. The conference is single track and is highly selective. Preceding the main session, there will be one day of tutorial presentations (November 29), and following it, there will be two days of focused workshops on topical issues at a Breckenridge (December 3-4). Major categories of accepted papers include the following topics: Algorithms and Architectures: supervised and unsupervised learning algorithms, model selection algorithms, active learning algorithms, feedforward and recurrent network architectures, localized basis functions, mixture models, belief networks, graphical models, Gaussian processes, factor analysis, topographic maps, combinatorial optimization, hybrid symbolic-subsymbolic systems. Applications: handwriting recognition, sequence analysis, expert systems, fault diagnosis, medical diagnosis, analysis of medical images, data analysis, database mining, information retrieval, network traffic, music processing, time-series prediction, financial analysis. Cognitive Science/Artificial Intelligence: perception and psychophysics, neuropsychology, cognitive neuroscience, development, conditioning, human learning and memory, attention, language, natural language, reasoning, spatial cognition, emotional cognition, conceptual representation, neurophilosophy, problem solving and planning. Implementations: analog and digital VLSI, optical neurocomputing systems, novel neurodevices, computational sensors and actuators, simulation tools. Neuroscience: neural encoding, spiking neurons, synchronicity, sensory processing, systems neurophysiology, neuronal development, synaptic plasticity, neuromodulation, dendritic computation, channel dynamics, experimental data relevant to computational issues. Reinforcement Learning and Control: exploration, planning, navigation, Q-learning, TD-learning, state estimation, dynamic programming, robotic motor control, process control, Markov decision processes. Speech and Signal Processing: speech recognition, speech coding, speech synthesis, auditory scene analysis, source separation, applications of hidden Markov models to signal processing, models of human speech perception, auditory modeling and psychoacoustics. Theory: computational learning theory, statistical physics of learning, information theory, prediction and generalization, regularization, Boltzmann machines, Helmholtz machines, decision trees, support vector machines, online learning, dynamics of learning algorithms, approximation and estimation theory, learning of dynamical systems, complexity theory. Visual Processing: image processing, image coding, object recognition, visual psychophysics, stereopsis, motion detection and tracking. For general inquiries or requests for registration material E-mail: nipsinfo at salk.edu or Fax: (619) 587-0417 Information can also be found on the NIPS web page. NIPS*99 Organizing Committee: General Chair, Sara Solla, Northwestern University; Program Chair, Todd Leen, Oregon Graduate Institute; Publications Chair, Klaus Mueller, GMD First; Tutorial Chair, Joachim Buhmann, University of Bonn; Workshops Co-Chairs, Sue Becker, McMaster University, and Rich Caruana, Carnegie Mellon University; Publicity Chair, Lee Giles, NEC Research Institute; Local Arrangements, Arun Jagota, University of California at Santa Cruz; Treasurer, Bartlett Mel, University of Southern California; Web Master, Doug Baker, Carnegie Mellon University; Government Liaison, Gary Blasdel, Harvard Medical School; Contracts, Steve Hanson, Rutgers University, and Gerry Tesauro, IBM. NIPS*99 Program Committee: Leon Bottou, AT&T Labs - Research; Gary Cottrell, University of California San Diego; Zoubin Ghahramani, University College London; Tommi Jaakkola, MIT; John Lazzaro, University of California Berkeley; Todd Leen, Oregon Graduate Institute (chair); John Moody, Oregon Graduate Institute; Barak Pearlmutter, University of New Mexico; Alexandre Pouget, University of Rochester; David Saad, Aston University; Lawrence Saul, AT&T Labs - Research; Sebastian Thrun, Carnegie Mellon University; Benjamin Van Roy, Stanford University; Paul Viola, MIT. - Please Post - __ C. Lee Giles / Computer Science / NEC Research Institute / 4 Independence Way / Princeton, NJ 08540, USA / 609-951-2642 / Fax 2488 www.neci.nec.com/~giles == From ishii at is.aist-nara.ac.jp Wed Oct 20 23:46:04 1999 From: ishii at is.aist-nara.ac.jp (Shin Ishii) Date: Thu, 21 Oct 1999 12:46:04 +0900 Subject: Paper Avialable Message-ID: <199910210346.MAA32143@axp27.aist-nara.ac.jp> Dear Connectionists, I am pleased to inform you that the following paper is available on my Web site: http://www.aist-nara.ac.jp/~ishii/publication.html We would greatly appreciate it, if you could give us any comments and suggestion. ----------------------------------------------------------------- Lambda-opt neural approaches to quadratic assignment problems Shin Ishii and Hirotaka Niitsuma to appear in Neural Computation Abstract: In this paper, we propose new analog neural approaches to combinatorial optimization problems, in particular, quadratic assignment problems (QAPs). Our proposed methods are based on an analog version of the lambda-opt heuristics, which simultaneously changes assignments for lambda elements in a permutation. Since we can take a relatively large lambda value, our new methods can achieve a middle-range search over possible solutions, and this helps the system neglect shallow local minima and escape from local minima. In experiments, we have applied our methods to relatively large-scale (N = 80-150) QAPs. Results have shown that our new methods are comparable to the present champion algorithms; for two benchmark problems, they are able to obtain better solutions than the previous champion algorithms. ----------------------------------------------------------------- Shin Ishii, Ph.D. Nara Institute of Science and Technology ATR Human Information Processing Research Laboratories From arthur.filippidis at dsto.defence.gov.au Thu Oct 21 03:36:56 1999 From: arthur.filippidis at dsto.defence.gov.au (Filippidis, Arthur) Date: Thu, 21 Oct 1999 17:06:56 +0930 Subject: Call for Papers, Special Fusion of Intelligent Systems Session at KES2000 Conference in UK, 30 Aug -1 Sept 2000 Message-ID: Special Session for KES2000 Conference. FUSION OF KNOWLEDGE-BASED INTELLIGENT SYSTEMS Architectures and Applications Organiser/Chair : Dr. Arthur Filippidis Please Circulate. Regards Dr. Arthur Filippidis System Concepts Group DSTO, Land Operations Division ph: (08) 8259 5442 fax: (08) 8259 5624 Fourth International Conference on Knowledge-Based Intelligent Engineering Systems & Allied Technologies (http://www.eng.bton.ac.uk/eee/research/kes2000/) 30 August-1 September 2000 - University of Brighton, Sussex, U.K. Special Session on : FUSION OF KNOWLEDGE-BASED INTELLIGENT SYSTEMS Architectures and Applications Call For Papers The special session aims to cover recent practical applications in the fusion of various Intelligent techniques such as, fuzzy, neural or genetic algorithms, or in combination with intelligent agents. The session also covers recent advances in knowledge-based multisensor or Information fusion architectures using a hybrid of various Intelligent techniques, in the practical areas such as for example landmine detection, target recognition, multispectral image fusion or other image fusion applications. Topics of the Special Session cover, but are not limited to: * Fusion of sensors using Intelligent techniques in areas such as target detection, landmines. * Combining the use of any Intelligent Agent technology / Fuzzy /Neural / and Genetic Algorithms to produce novel hybrid architectures used in practical applications * Multisensor Data Fusion architectures in practical applications using Intelligent techniques. * Image or Information Fusion using Intelligent techniques or Agent technology. * Multispectral fusion of Images or GIS Information using Knowledge-based techniques. INSTRUCTIONS FOR AUTHORS Interested authors must submit 2 copies of their paper to the Session Chair by post (prefered method), ( Directly to Dr. A. Filippidis, Postal Address below), or email (Word97 or earlier versions only) attachments also accepted. Submitted papers must comply to the Conference guidelines for submission (http://www.eng.bton.ac.uk/eee/research/kes2000/). All papers will be strictly refereed by at least three experts in the field. PUBLICATION The session papers will be published in the Conference Proceedings. IMPORTANT DATES Deadline for Submission: January 20th, 2000 Acceptance/Rejection Notification: March 20th, 2000 Camera-ready papers due April 10th, 2000 Session Chair: Dr. Arthur Filippidis Defence Science Technology Organisation Land Operations Division, P.O. Box 1500, Salisbury, South Australia, 5108, Australia. Email: arthur.filippidis at dsto.defence.gov.au ph: (08) 8259 5442 fax: (08) 8259 5624 From andre at icmc.sc.usp.br Fri Oct 22 04:46:39 1999 From: andre at icmc.sc.usp.br (Andre Carlos Ponce de Leon F. de Carvalho) Date: Fri, 22 Oct 1999 09:46:39 +0100 Subject: CFP for New Journal: IJCIA Message-ID: <3810246F.CBFA54AC@icmc.sc.usp.br> --------------------------------- Apologies for cross-posting --------------------------------- CALL FOR PAPERS ************************************************************* NEW JOURNAL BY IMPERIAL COLLEGE PRESS http://www.wspc.com/journals/ijcia/ijcia.html ************************************************************* INTERNATIONAL JOURNAL OF COMPUTATIONAL INTELLIGENCE AND APPLICATIONS (IJCIA) Editors-in-Chief: Andre de Carvalho, University of Guelph, Canada (until December 1999, University of Sao Paulo, Brazil) Antonio Braga, University Federal of Minas Gerais, Brazil Brijesh Verma, Griffith University, Australia Honorary Editors-in-Chief: Igor Aleksander, Imperial College, UK John Holland, University of Michigan and Santa Fe Institute, USA Lofti Zadeh, University of California at Berkeley, USA Associate Editors: Carme Torras, CSIC-UPC, Spain Harold Szu (NSWC, USA) Xin Yao, University of Birmingham, UK Witold Pedrycz, University of Alberta, Canada Editorial Board: A. Aamodt (NTNU, Norway), I. Aleksander (Imperial College, UK), J. A. Anderson (Brown University, USA), G. Cotrell (University of California San Diego, USA), Y. Dote (Muroran Institute of Technology, Japan), M. Fairhurst (University of Kent at Canterbury, UK), C. Lee Giles (NEC Research Institute, USA), M. Jamshidi (University of New Mexico, USA), J. Holland (University of Michigan and Santa Fe Institute, USA), Z. Liu (University of Melbourne, Australia), T. B. Ludermir (UFPE, Brazil), E. Mamdani (Imperial College, UK), J. Mulawka (Warsaw University of Technology, Poland), S. K. Pal (Statistical Institute of Calcutta, India), W. Pedrycz (University of Alberta, Canada), J. Shavlik (University of Wisconsin, USA), H. Szu (NSWC, USA), N. Kasabov (University of Otago, New Zealand), Y. Takefuji (Keio University , Japan), C. Torras (CSIC-UPC, Spain), G. Thimm (Nanyang Tech. Univ., Singapore), I. B. Turksen (University of Toronto, Canada), A. S. Weigend (New York University, USA), X. Yao, (University of Birmingham, UK), L. Zadeh, Univ. of Berkeley, USA Introduction: Computational Intelligence is a fastly expanding research field, attracting every year a large number of scientists, engineers and practitioners. Moreover, a growing number of companies is employing Computational Intelligence techniques to improve previous solutions and to deal with new problems. For a large number of applications, it has been realized that the combination of different techniques might lead to more efficient solutions. Aims and Scope: The International Journal of Computational Intelligence and Applications, IJCIA, is a refereed journal dedicated to the theory and applications of computational intelligence (artificial neural networks, fuzzy systems, genetic algorithms and hybrid systems). The main goal of the journal is to provide a vehicle to the scientific community and industry where ideas using two or more conventional and computational intelligence based techniques could be discussed. The journal welcomes original works in areas such as neural networks, fuzzy logic, evolutionary computation, pattern recognition, hybrid intelligent systems, symbolic machine learning, statistical models, image/audio/video compression and retrieval. The journal highly encourages new ideas, cobining two or more areas, such as neuro-fuzzy, neuro-symbolic, neuro-genetic, neuro-symbolic, neuro-pattern recognition, genetic-fuzzy, genetic-symbolic, fuzzy-symbolic, etc. Submission: Submitted papers must be in English and should not have been published elsewhere or being currently under consideration by any other journal. The original and four copies of the manuscript, not exceeding 10 pages, should be submitted to any one of the Editors-in-Chief. In addition to the hardcopies, authors should also send an electronic version of their papers in a disc (postscript or pdf formats are preferable). Authors should retain a copy of their submitted paper(s) as a guarantee for loss or damage. Addresses for submission : Andre C P L Ferreira de Carvalho Dept. of Ciencias de Computacao e Estatistica ICMC - Universidade de S?o Paulo Caixa Postal 668 CEP 13560-970 S?o Carlos, SP Brazil E-mail: andre at icmc.sc.usp.br ===> Moving in January 2000 to: <=== Dept. of Computing and Information Science University of Guelph Guelph, ON, N1G 2W1 Canada Tel.: +1 519 824-4120 Fax: +1 519 837-0323 E-mail: andre at snowhite.cis.uoguelph.ca Antonio de Padua Braga Dept. Engenharia Eletronica Campus da UFMG (Pampulha) Caixa Postal 209 30.161-970, Belo Horizonte, MG Brazil Tel.: +55 31 499 4869 (499 4848) Fax: +55 31 499 4850 E-mail: apbraga at eee.ufmg.br Brijesh Verma School of Information Technology Griffith University-Gold Coast Campus PMB 50, Gold Coast Mail Center QLD 9726 Australia Tel.: +61 755948592 Fax: +61 755948066 E-mail: b.verma at gu.edu.au "International Journal of Computational Intelligence and Applications will be distributed by World Scientific Press Company, Singapore." From abrowne at lgu.ac.uk Mon Oct 25 11:11:05 1999 From: abrowne at lgu.ac.uk (Tony Browne) Date: Mon, 25 Oct 1999 16:11:05 +0100 (GMT Daylight Time) Subject: Postdoc + PhD in UK Message-ID: Two research posts (Portsmouth/London) are available for projects in Biological Data Mining, one post-doctoral research fellow and one research studentship. A Postdoctoral Research Fellow is required for the project "Biological data mining: A comparison of neural network and symbolic techniques" funded by the joint BBSRC/EPSRC programme in bioinformatics. The Research Fellow will be based in the Centre for Molecular Design at the University of Portsmouth. This post requires either a graduate with a PhD in Computer Science, Cognitive Science, Mathematics or a related subject, with an interest in applying machine learning techniques in the field of bioinformatics; or a graduate with a PhD in the Life Sciences with a thorough understanding of bioinformatics and strong mathematical and programming skills. The successful candidate will be required to: * design, implement and test novel numerical algorithms in a high-level programming language * apply new and existing software to biological datasets * perform statistical analyses of the results The ideal candidate will possess knowledge in some of the following areas: * experience of developing neural networks in high level programming languages * strong mathematical and statistical skills (including Bayesian statistics) * an understanding of symbolic machine learning techniques * skills in Matlab, C/C++, Unix * an awareness of current developments in bioinformatics The starting salary for this post will be in the range stlg 17,606 - 19,203. A Postgraduate Research Student is required to perform research on using neural networks with biological and chemical datasets. The Research Student will be based in the School of Computing, Information Systems and Mathematics at London Guildhall University. This post requires either a graduate with a good first degree in Computer Science, Cognitive Science, Mathematics or a related subject, with an interest in applying machine learning techniques in the field of bioinformatics; or a graduate with a good first degree in the Life Sciences with a thorough understanding of bioinformatics and strong mathematical and programming skills. The successful candidate will be required to: * design, implement and test novel numerical algorithms in a high-level programming language * apply new and existing software to biological datasets * perform statistical analyses of the results A tax-free bursary is provided for this post, to a level commensurate with standard research council PHD studentships. The candidate will be encouraged to register for a PhD. As this studentship is internally funded some formal teaching will be required (typically up to 6 hours of laboratory supervision). Both of the above posts will be of 3 years duration, and will start as soon as possible. To apply please e-mail a CV (in text format) including the names (and e-mail addresses) of two referees to Dr. Antony Browne (abrowne at lgu.ac.uk) by 19th November 1999. ======================================================= Dr. Antony Browne School of Computing, Information Systems & Mathematics London Guildhall University 100 Minories London EC3 1JY, UK Tel: (+44) 0207 320 3106 Fax: (+44) 0207 320 1717 ======================================================= From bhaskar at crab.rutgers.edu Mon Oct 25 15:40:43 1999 From: bhaskar at crab.rutgers.edu (Bhaskar DasGupta) Date: Mon, 25 Oct 1999 15:40:43 -0400 (EDT) Subject: faculty position in neural computation Message-ID: <199910251940.PAA20520@bhaskar.rutgers.edu> Fellow Connectionists, Below is a faculty opening announcement in our department that would be relevant to the neural network and the learning theory community. It will also appear (sometime in future) in CACM, IEEE Computer, and the Computing Research News. Please feel free to pass it to any other interested party. My apologies if you received this multiple times or if this announcement was not relevant to you. Best regards, Bhaskar DasGupta ------------------------------------------------------------------------------ Bhaskar DasGupta Department of Computer Science Email: bhaskar at crab.rutgers.edu 319 Business & Science Building Phone: 856-225-6439 (office, in Rutgers) 3rd & Penn Street 732-445-4580 (office, in DIMACS) Rutgers University Fax: 856-225-6624 (office) Camden, NJ 08102-1656, USA URL:http://crab.rutgers.edu/~bhaskar ------------------------------------------------------------------------------ ================================================================================ Rutgers University - Camden Department of Computer Science The Department of Computer Science at Rutgers University - Camden invites applications for two tenure-track positions at the assistant professor level beginning September 2000. Candidates in all areas of computer science are encouraged to apply, particularly those with research interests in graphics and visualization, computational geometry, artificial neural networks, machine learning, networking, distributed and mobile computing, and parallel computing. A Ph.D. in Computer Science is required. Salary is contingent upon qualifications. Rutgers University stands in the top 5% of AAUP rankings of university salary and benefits. Applicants should send a curriculum vita, statement of research, statement of teaching interests/experience, and three letters of reference to: Chair, Faculty Search Committee Department of Computer Science Rutgers University Camden, NJ 08102 Review of applications will begin immediately and will continue until the positions are filled. Rutgers University is an Equal Opportunity and Affirmative Action Employer. From baolshausen at ucdavis.edu Mon Oct 25 20:54:19 1999 From: baolshausen at ucdavis.edu (Bruno Olshausen) Date: Mon, 25 Oct 1999 17:54:19 -0700 Subject: Institute for Theoretical Dynamics, UC Davis Message-ID: <3814FBBB.CB843FC8@ucdavis.edu> Dear Colleagues, The University of California, Davis, and the National Science Foundation have established a multidisciplinary training group, "Nonlinear Dynamics in Biology". The program is designed to provide graduate students with research training that will allow them to address questions in biology using computational and mathematical modelling. Specifically, our trainees focus on complex processes in cell biology, neuroscience, biofluiddynamics, ecology, and population biology that involve nonlinear dynamics. The curriculum includes training in computational, mathematical, and experimental methods. Students are mentored by faculty trainers, postdoctoral researchers, and more advanced graduate students in both formal and informal settings. Financial support is available for all graduate trainees. Admission to the graduate training program requires strong interest in biological research that involves nonlinear dynamics, acceptance to a graduate program at UC Davis, and a year-long course in calculus. We strongly encourage additional coursework in mathematics, computer science, and biology. Please alert students who might be interested in this program to the training grant web site, http://www.itd.ucdavis.edu/rtg/, for additional information and application materials. You also may contact our academic coordinator, Carole Hom (clhom at ucdavis.edu; 530-754-9117) should you or your students have questions not addressed in the information on the web site. In addition, we seek postdoctoral fellows in biofluiddynamics, cell or neurobiology, and population biology or ecology. Application materials for postdocs also can be found on the web page. Thank you for your assistance. -- Bruno A. Olshausen Phone: (530) 757-8749 Center for Neuroscience Fax: (530) 757-8827 UC Davis Email: baolshausen at ucdavis.edu 1544 Newton Ct. WWW: http://redwood.ucdavis.edu Davis, CA 95616 From luettin at idiap.ch Tue Oct 26 12:50:23 1999 From: luettin at idiap.ch (Juergen Luettin) Date: Tue, 26 Oct 1999 18:50:23 +0200 Subject: Postdocs/PhDs in Multimodal Biometrics and Multimedia Retrieval Message-ID: <199910261650.SAA21967@montfort.idiap.ch> POSTDOC and PHD POSITIONS in MULTIMODAL BIOMETRICS and MULTIMEDIA RETRIEVAL at IDIAP, Switzerland Applications are invited for several Post-Doctoral positions and PhD research studentships in the Computer Vision Group at IDIAP to work in one of the following research areas (with possible research topics): Video Indexing/Retrieval - adaptation of video content analysis methods - combination and integration of audio, video, and text information - high level video structure modelling - video abstraction/summary generation, browsing, GUI, MPEG7 Audio Indexing/Retrieval - adaptation of speech recognition systems to video (acoustic models, lexicon, language models, topic) - audio segmentation (speech, music, audio events) - speaker change detection - document indexing and retrieval, MPEG7 Text Segmentation and Recognition in Videos - detection, segmentation and recognition of text embedded in video - investigation of image properties (texture, colour, shape) - use of temporal information - integration of detection/segmentation algorithms with text recognition engine Face Analysis - face detection (in complex backgrounds, under variant pose, occlusion) - face recognition (identification/verification, robustness issues, confidence estimation, evidence integration) Speaker Verification - client model adaptation - decision strategy - robustness to microphone/channel characteristics - discriminant analysis Classifier Combination - classifier confidence estimation - classifier combination techniques - evidence integration - application to multimodal biometrics, video retrieval The successful candidates will participate in the following important and highly competitive European and national projects: o BANCA: Biometric Access Control for Networked and e-Commerce EU project, lead by Matra Nortel Communication http://www.idiap.ch/vision/banca.html o ASSAVID: Automatic Segmentation and Semantic Annotation of Sports Videos EU project, lead by Sony UK http://www.idiap.ch/vision/assavid.html o VOCR: Text Segmentation and Recognition for Video Retrieval Swiss National Science Foundation project http://www.idiap.ch/vision/vocr.html Several tasks will be performed in close collaboration with other project partners including universities, industries, and end-users. REQUIREMENTS Postdoctoral candidates should possess a relevant PhD degree and doctoral students an M.S. degree (or equivalent) in Computer Science or a related area and should have a strong background in mathematics, signal processing, and pattern recognition. ABOUT IDIAP IDIAP is a semi-private research institute, that conducts basic and applied research in artificial intelligence. The main research activities are in machine learning, speech processing, and computer vision. Research in the rapidly growing computer vision group focuses on topics in multimedia processing including multimodal biometrics, multimodal speech recognition, and document analysis and recognition (See http://www.idiap/vision). Doctoral students at IDIAP are usually registered at a Swiss university or Federal Institute of Technology, often EPFL, that awards the PhD degree. LOCATION IDIAP is located in the town of Martigny in Valais, a scenic region in the south of Switzerland, surrounded by some of the highest mountains in Europe which offer some of the best skiing, hiking, and climbing. It is within close proximity to Lausanne, and lake Geneva, and centrally located for travel to other parts of Europe. HOW TO APPLY Prospective candidates should send a detailed CV, including 2 names and addresses of referees, research interest, and earliest possible starting date, preferably by email to Dr. Juergen Luettin Head, Computer Vision Group IDIAP, CP 592, 1920 Martigny, Switzerland Email: Luettin at idiap.ch Tel: +41 27 721 77 27 Fax: +41 27 721 77 12 From ken at phy.ucsf.EDU Wed Oct 27 05:56:53 1999 From: ken at phy.ucsf.EDU (Ken Miller) Date: Wed, 27 Oct 1999 02:56:53 -0700 (PDT) Subject: Paper available: Review of Orientation Selectivity Development and Models Message-ID: <14358.52325.953582.526768@coltrane.ucsf.edu> The following paper is now available at ftp://ftp.keck.ucsf.edu/pub/ken/or-review.ps.gz (compressed postscript) or http://www.keck.ucsf.edu/~ken (click on 'Publications') This is a preprint of an article that appeared as Journal of Neurobiology 41:44-57 (1999): http://www3.interscience.wiley.com/cgi-bin/abstract?ID=66000451 ------------------------------ Is the Development of Orientation Selectivity Instructed by Activity? Kenneth D. Miller, Ed Erwin and Andrew Kayser Dept. of Physiology, UCSF ABSTRACT: Is the development of orientation selectivity in visual cortex instructed by the patterns of neural activity of input neurons? We review evidence as to the role of activity, review models of activity-instructed development, and discuss how these models can be tested. The models can explain the normal development of simple cells with binocularly matched orientation preferences, the effects of monocular deprivation and reverse suture on the orientation map, and the development of a full intracortical circuit sufficient to explain mature response properties including the contrast-invariance of orientation tuning. Existing experiments are consistent with the models, in that (1) selective blockade of ON-center ganglion cells, which will degrade or eliminate the information predicted to drive development of orientation selectivity, in fact prevents development of orientation selectivity; and (2) the spontaneous activities of inputs serving the two eyes are correlated in the lateral geniculate nucleus at appropriate developmental times, as was predicted to be required to achieve binocular matching of preferred orientations. However, definitive tests remain to be done to (1) firmly establish the instructive rather than simply permissive role of activity and (2) determine whether the retinotopically- and center-type-specific patterns of activity predicted by the models actually exist. We conclude by critically examining alternative scenarios for the development of orientation selectivity and maps, including the idea that maps are genetically pre-specified. Ken Kenneth D. Miller telephone: (415) 476-8217 Dept. of Physiology fax: (415) 476-4929 UCSF internet: ken at phy.ucsf.edu 513 Parnassus www: http://www.keck.ucsf.edu/~ken San Francisco, CA 94143-0444 From mw at stat.Duke.EDU Wed Oct 27 09:27:48 1999 From: mw at stat.Duke.EDU (Mike West) Date: Wed, 27 Oct 1999 09:27:48 -0400 Subject: No subject Message-ID: <19991027092748.E26398@isds.duke.edu> ------------------------------------------------------------------------- ------------------------------------------------------------------------- MITCHELL PRIZE 2000: ANNOUNCEMENT AND SOLICITATION http://www.stat.duke.edu/sites/mitchell.html The Mitchell Prize is awarded in recognition of an outstanding paper that describes how a Bayesian analysis has solved an important applied problem. The 2000 Prize includes an award of $1000 and a commemorative plaque, and will be announced and presented at the ISBA 2000 meeting in Crete (May 28-June 1 2000). The Mitchell Prize is named for Toby J. Mitchell and was established by his friends and colleagues following his death from leukemia in 1993. Toby was a Senior Research Staff Member at Oak Ridge National Laboratory throughout his career, will leaves of absence spent at the University of Wisconsin and at the National Institute of Environmental Health Sciences. Toby won the Snedecor Award in 1978 (with co-author Bruce Turnbull), made incisive contributions to statistics, especially in biometry and engineering applications, and was a marvelous collaborator and an especially thoughtful scientist. Toby was a dedicated Bayesian, hence the focus of the prize. This is the fourth Mitchell Prize, the first three having been awarded in 1994, 1997 and 1999. Since 1999 the Prize is awarded annually under the cosponsorship of the ASA Section on Bayesian Statistical Science (SBSS), the International Society for Bayesian Analysis (ISBA), and the Mitchell Prize Founders' Committee. The awarding of the Mitchell Prize is governed by the Mitchell Prize charter, established in 1999 (and available at the Mitchell Prize web site, noted above). Under this charter, the sponsors annually establish a selection committee; the 2000 Prize selection committee members are Gary Koop, Henry Wynn and Mike West (chair). To be eligible for the 2000 Prize, a paper will either have appeared in a refereed journal or refereed conference proceedings since January 1 1998, or be scheduled for future publication in a refereed outlet. Candidate papers will be accepted from nominators and from authors. In reviewing submissions, emphasis will be placed on evidence that the application has truly benefited from a Bayesian analysis respecting the individual character of the problem at hand. There is no restriction as to approach taken, except that it be Bayesian in some sense, and that it carefully and appropriately justifies models, priors and methodologies adopted. To be considered for this year's Prize, please submit the following: * FOUR reprints or copies of the manuscript * A cover letter, with two copies, containing the following elements: -- A brief statement of the impact of the work -- Contact information for the authors and nominator (if not an author) -- Full email and postal addresses, plus telephone numbers, for TWO individuals who can be contacted for an evaluation of the importance of the work in the applied field. The named individuals should be experts in the applied field in question, but must not be either statisticians or coauthors/collaborators of those named on the submission. Submissions should be mailed to Mike West, Mitchell Prize Selection Committee Chair Institute of Statistics & Decision Sciences Duke University Durham, NC 27708-0251 USA Entries must be received at this address by JANUARY 31st 2000 in order to receive consideration. Visit the web site listed above to learn more about the Mitchell Prize and the sponsoring organizations. ------------------------------------------------------------------------ ------------------------------------------------------------------------ From Frederic.Alexandre at loria.fr Thu Oct 28 10:30:37 1999 From: Frederic.Alexandre at loria.fr (Frederic Alexandre) Date: Thu, 28 Oct 1999 16:30:37 +0200 Subject: PostDoc Position at LORIA-INRIA France Message-ID: <38185E0D.6257@loria.fr> LORIA/INRIA Lorraine computer science laboratory in Nancy, France One postdoctoral position is available for developping image processing systems in the framework of computational neurosciences applied to robotics, in Nancy, France, from January to December 2000. Our team: the CORTEX team is developping connectionist models, inspired from biology, for perception, reasoning and autonomous behavior. Belonging to a computer science lab, our main goal is to propose effective models for robotics, speech and image processing. The position will consist in helping PhD students to build the visual processing system that have to be interfaced to their biologically inspired models. This system will be itself biologically inspired or will simply include classical image processing algorithms. The work thus implies developping an interface, for a Koala autonomous robot, between a video camera and our cortically inspired models. The computer scientist candidate: he/she should be well-trained in image processing and connectionist modeling. Experience in robotics, autonomous behavior and neurosciences would be highly appreciated. For further information, contact: ---------------------------------------------------------------------------- Frederic ALEXANDRE Tel: (+33/0) 3 83 59 20 53 INRIA-Lorraine/LORIA-CNRS Fax: (+33/0) 3 83 41 30 79 BP 239 E-mail: falex at loria.fr 54506 Vandoeuvre-les-Nancy Cedex http://www.loria.fr/~falex FRANCE ---------------------------------------------------------------------------- From maass at igi.tu-graz.ac.at Fri Oct 29 17:04:49 1999 From: maass at igi.tu-graz.ac.at (Wolfgang Maass) Date: Fri, 29 Oct 1999 23:04:49 +0200 Subject: Neural nets with dynamic synapses and nonlinear filters Message-ID: <381A0BF1.B4117AF6@igi.tu-graz.ac.at> The following paper is now online available: Neural Systems as Nonlinear Filters Wolfgang Maass and Eduardo D.Sontag Technische Univ. Graz Rutgers University Abstract: Experimental data show that biological synapses behave quite differently from the symbolic synapses in all common artificial neural network models. Biological synapses are dynamic, i.e., their ``weight'' changes on a short time scale by several hundred percent in dependence of the past input to the synapse. In this article we address the question how this inherent synaptic dynamics-- which should not be confused with long term ``learning'' -- affects the computational power of a neural network. In particular we analyze computations on temporal and spatio-temporal patterns, and we give a complete mathematical characterization of all filters that can be approximated by feedforward neural networks with dynamic synapses. It turns out that even with just a single hidden layer such networks can approximate a very rich class of nonlinear filters: all filters that can be characterized by Volterra series. This result is robust with regard to various changes in the model for synaptic dynamics. Our characterization result provides for all nonlinear filters that are approximable by Volterra series a new complexity hierarchy which is related to the cost of implementing such filters in neural systems. The article will appear in Neural Computation. It is online available as # 107 from http://www.tu-graz.ac.at/igi/maass/#Publications and from http://www.math.rutgers.edu/~sontag/FTP_DIR/spiking.ps.gz From reggia at cs.umd.edu Fri Oct 29 12:48:09 1999 From: reggia at cs.umd.edu (James A. Reggia) Date: Fri, 29 Oct 1999 12:48:09 -0400 (EDT) Subject: postdoc position, computational neuroscience and language Message-ID: <199910291648.MAA28723@avion.cs.umd.edu> Post-doctoral Fellowships in the Cognitive Neuroscience of Language and its Disorders Two-year National Research Service Award fellowships are available at the University of Maryland School of Medicine, in Baltimore, Maryland. Training opportunities will provide experience in the application of contemporary research methods (including computational modeling, cognitive neuropsychology, event-related potentials and functional neuroimaging) to the topic of normal and disordered language processing. Applicants with doctoral degrees in related basic science areas (computer sciene, neuroscience, linguistics, cognitive psychology, etc.) and clinical disciplines (speech/language pathology; clinical neuropsychology) are invited to apply. Applicants must be U.S. citizens or permanent residents to be considered, under the terms of the NRSA program. Inquiries may be directed to Rita Berndt at rberndt at umaryland.edu or to Jim Reggia at reggia at cs.umd.edu . To apply, send HARD COPIES of your C.V., the names and addresses of three referees, and a statement of research interests and career goals to: Rita S. Berndt, Ph.D. Department of Neurology University of Maryland School of Medicine 22 South Greene Street Baltimore, Maryland 21201 USA From Nigel.Goddard at ed.ac.uk Mon Oct 25 22:38:42 1999 From: Nigel.Goddard at ed.ac.uk (Nigel Goddard) Date: Tue, 26 Oct 1999 03:38:42 +0100 Subject: Jobs in Neural Simulation Message-ID: <38151432.1BEE906A@ed.ac.uk> POSTDOCTORAL RESEARCH FELLOW & RESEARCH PROGRAMMER IN NEURAL SIMULATION METHODS The NEOSIM project, funded by NIMH and NSF, is developing simulation environments and associated software tools for modelling of brain processes. The project has computer science and neuroscience research goals, and aims to develop distributable, portable, parallel software for the computational neuroscience community. We seek a postdoctoral Research Fellow and a Research Programmer to augment the existing multidisciplinary, multinational team. Both posts may require some travel to other EU countries or the US. The Research Fellow will conduct original research related to brain modeling, brain data analysis or parallel simulation techniques and will contribute design, documentation, programming or other relevant expertise to the software development goals of the project. The Research Programmer will be involved in design, documentation, implementation, testing, dissemination, maintenance and development of the NEOSIM framework and will also be responsible for the installation and maintenance of software packages on the research computers, and may be involved in the specification, installation and system administration of research computers including small parallel platforms. Tenable for up to 3 years subject to annual renewal and further renewal beyond 3 years subject to success in attracting further funding. Salary scale: ?16,286 - ?24,479 p.a. Further details are available at http://anc.ed.ac.uk/neosim/rafurthpart.html Informal enquiries to: Dr. Nigel Goddard (Nigel.Goddard at ed.ac.uk) Sites: http://anc.ed.ac.uk/neosim, http://www.informatics.ed.ac.uk, http://www.personnel.ed.ac.uk/recruit.htm Further particulars, including details of the application procedure should be obtained from the Personnel Department, 1 Roxburgh Street, Edinburgh EH8 9TB, tel: +44 131-650-2511 (24 hour answering service); or see the web site above Closing Date: 24th November 1999. Interviews expected 9th-13th December. -- ======================= Dr. Nigel Goddard Institute for Adaptive and Neural Computation Division of Informatics University of Edinburgh 5 Forrest Hill Edinburgh EH1 2QL Scotland Telephone: +44 131 650 3087 email: Nigel.Goddard at ed.ac.uk web: http://anc.ed.ac.uk/~ngoddard Fax (paper): +44 131 650 6899 eFAX (email): +1 603 698 5854 =======================