From nnsp02 at neuro.kuleuven.ac.be Mon Jul 1 06:52:52 2002 From: nnsp02 at neuro.kuleuven.ac.be (Neural Networks for Signal Processing 2002) Date: Mon, 01 Jul 2002 12:52:52 +0200 Subject: 2002 IEEE International Workshop on Neural Networks for Signal Processing; Early Registration Deadline: July 31, 2002 Message-ID: <3D203484.734B21B3@neuro.kuleuven.ac.be> ------------------------------------------------------------------------- 2002 IEEE International Workshop on Neural Networks for Signal Processing September 4-6, 2002 Martigny, Valais, Switzerland Early registration deadline: July 31, 2002 ------------------------------------------- http://isp.imm.dtu.dk/nnsp2002 The twelfth in a series of IEEE NNSP workshops will be held in Martigny, Switzerland at the H=F4tel du Parc - Centre du Congres. The workshop will feature a strong technical program complemented by a series of exciting plenary talks, a special session on bioinformatics and a tutorial on "Nonlinear Adaptive Filtering for MIMO Multi-Antenna Wireless Communications" given by Prof. S.Y. Kung. All will be included in the registration package as well as special events such as a banquet in an old monastery and reception at the Gianadda museum. Please refer to the workshop web site for more information and note the early registration deadline of July 31, 2002. Plenary Talks: Faking unlabeled data for geometric regularization by Yoshua Bengio, University of Montreal, Canada Occam's Razor and Infinite Models by Zoubin Ghahramani, University College London, UK Hidden Markov models of proteins and DNA by Anders Krogh, Copenhagen University, Denmark From Peter.Flach at bristol.ac.uk Tue Jul 2 14:09:06 2002 From: Peter.Flach at bristol.ac.uk (Peter A. Flach) Date: Tue, 2 Jul 2002 19:09:06 +0100 Subject: MSc in Machine Learning and Data Mining at Bristol, UK Message-ID: *** please post and distribute as appropriate *** The Computer Science Department at the University of Bristol has been running an Advanced Computing MSc in Machine Learning and Data Mining since 1998. We are still accepting applications for enrolment in October 2002. See the URLs below for more information on the course and on how to apply. MSc Admissions: http://www.cs.bris.ac.uk/Admissions/msc.html MLDM MSc course: http://www.cs.bris.ac.uk/Teaching/MachineLearning/ Machine Learning research: http://www.cs.bris.ac.uk/Research/MachineLearning/ --Peter -- +------------------------------------+------------------------------------+ Peter A. Flach, Reader in Machine Learning Peter.Flach at bristol.ac.uk Dept. of Computer Science, University of Bristol tel. +44 117 954 5162 Merchant Venturers Building, Woodland Road fax +44 117 954 5208 Bristol BS8 1UB, United Kingdom http://www.cs.bris.ac.uk/~flach/ +------------------------------------+------------------------------------+ From dgw at MIT.EDU Fri Jul 5 11:46:21 2002 From: dgw at MIT.EDU (David Weininger) Date: Fri, 05 Jul 2002 11:46:21 -0400 Subject: book announcement--Rao Message-ID: <2002070511462130675@outgoing.mit.edu> I thought readers of the Connectionists List might be interested in this book. For more information, please visit http://mitpress.mit.edu/0262182246/ Thank you! Best, David Probabilistic Models of the Brain Perception and Neural Function edited by Rajesh P. N. Rao, Bruno A. Olshausen, and Michael S. Lewicki Neurophysiological, neuroanatomical, and brain imaging studies have helped to shed light on how the brain transforms raw sensory information into a form that is useful for goal-directed behavior. A fundamental question that is seldom addressed by these studies, however, is why the brain uses the types of representations it does and what evolutionary advantage, if any, these representations confer. It is difficult to address such questions directly via animal experiments. A promising alternative is to use probabilistic principles such as maximum likelihood and Bayesian inference to derive models of brain function. This book surveys some of the current probabilistic approaches to modeling and understanding brain function. Although most of the examples focus on vision, many of the models and techniques are applicable to other modalities as well. The book presents top-down computational models as well as bottom-up neurally motivated models of brain function. The topics covered include Bayesian and information-theoretic models of perception, probabilistic theories of neural coding and spike timing, computational models of lateral and cortico-cortical feedback connections, and the development of receptive field properties from natural signals. Rajesh P. N. Rao is Assistant Professor in the Department of Computer Science and Engineering at the University of Washington. Bruno A. Olshausen is Associate Professor in the Department of Psychology and the Center for Neuroscience at the University of California, Davis. Michael S. Lewicki is Assistant Professor in the Department of Computer Science and the Center for the Neural Basis of Cognition at Carnegie Mellon University. 8 x 10, 345 pp., cloth, ISBN 0-262-18224-6 Neural Information Processing series A Bradford Book ______________________ David Weininger Associate Publicist The MIT Press 5 Cambridge Center, 4th Floor Cambridge, MA 02142 617 253 2079 617 253 1709 fax http://mitpress.mit.edu From jls at cs.man.ac.uk Fri Jul 5 05:05:41 2002 From: jls at cs.man.ac.uk (Jon Shapiro) Date: Fri, 5 Jul 2002 10:05:41 +0100 Subject: Lectureship in Robotics; Manchester University Message-ID: <20020705100541.M1699@mortlach.cs.man.ac.uk> University of Manchester Department of Computer Science Lecturer in Robotics The Computer Science Department at the University of Manchester is seeking to appoint a lecturer in Robotics. The Department has recently been given a 5* rating in the Research Assessment Exercise. The Department has built up one of the UK leading mobile robotics labs, supported by a well-trained technical support staff. We currently teach both an undergraduate and an MSc course in mobile robotics which the successful candidate will be expected to take over. It is expected that the appointee will join a strong Artificial Intelligence group and contribute to or complement the current activities in machine learning, spatial and temporal reasoning, and computation of language. There will also be the opportunity to collaborate with other research groups within the Department. Of particular interest are the following: - the Imaging Science and Biomedical Engineering group, who are interested in medical robotics and computer-aided surgery; - the Mobile Systems Architecture group, who need a platform for research into terrain mapping and geolocation, using radio communication signals in a mobile Wireless LAN; - the Advanced Interfaces group, who are interested in using video footage captured by a robot camera to reverse engineer buildings. The successful candidate will have a track record of publications in these or related areas. More information about current robotics activity in the Department can be found at http://www.cs.man.ac.uk/robotics/ Annex A Particulars of Appointment - Lecturer in Robotics, Department of Computer Science Job Description The Lecturer's primary duties will lie in teaching, research, and administration. During the probationary period, the Lecturer will expect to be granted some relief from standard teaching and administrative duties in order to develop their research capabilities. Research The appointee will join a strong Artificial Intelligence group and will be expected to contribute to or complement the current research activities in machine learning, spatial and temporal reasoning, and computation of language. The appointee will undertake research in their specialised area of robotics in one of the UK leading mobile robotics labs, built up by the Department over a number of years and will be supported by a well-trained technical support staff. There will also be the opportunity to collaborate with other research groups within the Department. Of particular interest are the following: - the Imaging Science and Biomedical Engineering group, who are interested in medical robotics and computer-aided surgery; - the Mobile Systems Architecture group, who need a platform for research into terrain mapping and geolocation, using radio communication signals in a mobile Wireless LAN; - the Advanced Interfaces group, who are interested in using video footage captured by a robot camera to reverse engineer buildings. The successful candidate will be expected to produce high quality publications in these or related areas, attending relevant conferences and workshops and disseminating recent research results whenever possible. The appointee will also be expected to make applications for research grants and funding from relevant bodies. More information about current robotics activity in the Department can be found at http://www.cs.man.ac.uk/robotics/ Teaching The successful candidate will be expected to undertake teaching both at undergraduate and postgraduate level and to contribute to the development of Computer Science teaching. The appointee will take over the teaching of the current undergraduate and MSc modules in mobile robotics. The appointee will be expected to develop teaching material and learning experiences for students in the light of current educational practice, in particular by attending the University's training course for newly appointed lecturers and to participate in the planning and development of courses within the framework of Departmental and Faculty committees. Other teaching duties may include supervising laboratory and examples classes, conducting tutorials and supervise undergraduate and MSc projects. The appointee may also be required to undertake the supervision of research students. Administration The appointee will be expected to undertake Departmental administrative duties as agreed with the Head of Department. Person Specification Applicants for the lectureship will be expected to have: - a good honours degree in Computer Science (and preferably a PhD) - several years experience in a research environment, either in industry or academia - a strong publication record in the relevant area (see Job Description above) or, if coming from industry, an equivalent indication of output - the ability and willingness to teach in their specialised areas and to contribute to core computer science teaching - the ability and willingness to undertake normal administrative duties - ability to work as part of a team - good communication skills - preferably some research experience as part of a multidisciplinary team ----------------------------------------------------------------------- Jonathan Shapiro Computer Science Dept Room: 2.34 University of Manchester Phone: 44-(0)161 275 6253 Oxford Road Fax: 44-(0)161 275 6204 Manchester M13 9PL E-mail: jls at cs.man.ac.uk United Kingdom. FTP site: ftp.cs.man.ac.uk in directory /pub/ai/jls WWW site: http://www.cs.man.ac.uk/ai/jls/jls.html From ingber at ingber.com Fri Jul 5 12:33:56 2002 From: ingber at ingber.com (Lester Ingber) Date: Fri, 5 Jul 2002 12:33:56 -0400 Subject: Financial Engineer Position Message-ID: <20020705163356.GA18280@alumnus> If you have very strong credentials for the position described below, please email your resume to: Lester Ingber Director R&D Financial Engineer A disciplined, quantitative, analytic individual proficient in prototyping and coding (such as C/C++, Maple/Mathematica, or Visual Basic, etc.) is sought for financial engineering/risk:reward optimization research position with established Florida hedge fund (over two decades in the business and $1 billion in assets under management). A PhD in a mathematical science, such as physics, statistics, math, or computer-science, is preferred. Hands-on experience in the financial industry is required. Emphasis is on applying state-of-the-art methods to financial time-series of various frequencies. Ability to work with a team to transform ideas/models into robust, intelligible code is key. Salary: commensurate with experience, with bonuses tied to the individual's and the firm's performance. Selection Process All applicants will be reviewed, and a long list will be generated for phone interviews. Other applicants will not be contacted further. From these phone interviews, a short list will be generated for face-to-face interviews. Our estimated date for final selection will be September to October 2002. Start date for this position may range anywhere from immediately to six months thereafter, depending on both the candidate's and the firm's needs. All available information is posted on http://www.ingber.com/open_positions.html -- Prof. Lester Ingber ingber at ingber.com ingber at alumni.caltech.edu www.ingber.com www.alumni.caltech.edu/~ingber From pfbaldi at ics.uci.edu Sat Jul 6 11:59:54 2002 From: pfbaldi at ics.uci.edu (Pierre Baldi) Date: Sat, 6 Jul 2002 08:59:54 -0700 Subject: Postdoctoral positions in Bioinformatics and Machine Learning, Univeristy of California, Irvine Message-ID: <000001c22506$2b13bae0$be04c380@time-slice.ics.uci.edu> Several NIH-supported postdoctoral positions in the areas of Bioinformatics and Machine Learning are available in the Department of Information and Computer Science (www.ics.uci.edu) and the Institute for Genomics and Bioinformatics (www.igb.uci.edu) at the University of California, Irvine. Areas of particular interest: protein structure/function prediction, analysis of high-throughput array data (e.g. DNA microarray data), gene regulation, systems biology, and all areas of machine learning. Prospective candidates should apply with a cover letter, CV, and names and email addresses of 2-3 referees to be sent, preferably by email, to: pfbaldi at ics.uci.edu. The positions are available immediately and the duration of the appointments are typically 2 years with possibility of renewal. Relevant faculty in the department include: P. Baldi, D. Kibler, R. Lathrop, E. Mjolsness, and P. Smyth. From: esann To: "Connectionists at cs.cmu.edu" References: From bogus@does.not.exist.com Mon Jul 8 11:42:56 2002 From: bogus@does.not.exist.com () Date: Mon, 8 Jul 2002 17:42:56 +0200 Subject: Call for special sessions: ESANN'2003 European Symposium on Artificial Neural Networks Message-ID: ---------------------------------------------------- | | | ESANN'2003 | | | | 11th European Symposium | | on Artificial Neural Networks | | | | Bruges (Belgium) - April 23-24-25, 2003 | | | | First announcement & call for special sessions | ---------------------------------------------------- The ESANN'2003 conference European Symposium on Artificial Neural Networks) will take place in Bruges (Belgium), on April 23-25, 2003. Please have a look to http://www.dice.ucl.ac.be/esann for details. Proposals for special sessions are solicited. Details on special sessions (what are the role, advantages and duties of special session organizers) are available at http://www.dice.ucl.ac.be/esann/callforspecialsessions.htm. We try as much as possible to avoid multiple sendings of this call for papers; however please apologize if you receive this e-mail twice, despite our precautions. You will find below a short version of this call for special sessions. What are special sessions ? --------------------------- A "special session" is a session that focuses on a particular topic in the artificial neural networks field. It is organized by a renowned scientist who solicits (but does not "invite") potential contributors. During the conference, the session begins with an introduction (or tutorial) to the session topic by the session organizer, making it easier for the participants to follow the subsequent talks. The following are examples of special sessions organized during ESANN'98 to 2002 conferences: Radial basis networks Neural networks for control Self-organizing maps for data analysis ANN for speech processing Cellular neural networks ANN for the processing of facial information Adaptive computation of data structures Remote sensing spectral image analysis Support vector machines Information extraction using unsupervised neural networks Spiking neurons Artificial neural networks and robotics Neural networks in medicine Time series prediction ANN for energy management systems Dedicated hardware implementations: perspectives on systems and applications Novel neural transfer functions ANN and evolutionary/genetic algorithms: hybrid approaches ANN in finance ANN and early vision processing Perspectives on Learning with Recurrent Networks Representation of high-dimensional data Neural Network Techniques in Fault Detection and Isolation Hardware and Parallel Computer Implementations of Neural Networks Exploratory Data Analysis in Medicine and Bioinformatics Neural Networks and Cognitive Science Who can organize special sessions ? ----------------------------------- Any scientist in the neural networks field is welcome to submit his/her proposal to organize a special session. He/she should first check is he/she is able to contact enough potential contributors in order to organize a sound session. How to organize special sessions ? ---------------------------------- We see the role of a session organizer as follows: to contact about 7-10 potential authors among specialists in the field and to ask them to submit a paper to the conference; to make sure that each paper is evaluated by 3 reviewers (including the session organizer); according to the reviews, to decide the final contents of the session, in agreement with the program committee; to chair the session during the conference; to present a tutorial (typically 30 minutes) at the beginning of the session. This tutorial should both introduce the topic to non-specialists, and make the link between the state-of-the-art and the papers presented in the session. to write a paper for the proceedings, with the contents of the tutorial. The following aspects must be taken into account: The session can include oral and poster presentations. Posters presentations are introduced by a 2-minutes spotlight during the oral session. There should be between 4 and 6 oral presentations during the session (including the introduction/tutorial), and between 0 and 6 posters. Soliciting contributors does NOT mean that their contributions are automatically accepted. Solicited contributions are submitted to a review process and will be rated according to their scientific value. This must be clearly explained by the session organizer when he/she first contacts the potential contributors. Soliciting contributors does NOT mean too that these contributors are invited to the conference. Even in case of acceptation of their submission, they will have to register to the conference and pay the registration fee, as any other participant. This must be clearly explained too by the session organizer when he/she first contacts the potential contributors. What are the advantages for special session organizers ? -------------------------------------------------------- Session organizers benefit from a 50% reduction of their own registration fees. However, travel and subsistence expenses must be taken in charge by the session organizers. Their article in the ESANN proceedings are limited to 12 pages instead of 6 pages. Depending on public funding obtained to organize the conference, the hotel costs of session organizers could be taken in charge. Nevertheless, we will not have the confirmation of this possibility before having to decide which special sessions will be organized. Those who submit a proposal to organize a special session must be aware that they could have to take in charge their hotel expenses. Deadlines --------- August 1, 2002: submission of proposals to organize a special session August 15, 2002: selection of special sessions for ESANN'2002 September 1, 2002: availability of the ESANN'2002 call for papers on this Web site end of September, 2002: contacts (by session organizer) with potential contributors December 6, 2002: deadline submission of papers (including special sessions) around December 15, 2002: dispatching of submissions to reviewers January 17, 2003: end of review process January 29, 2003: decision on the contents of the special sessions February 6, 2003: notification of acceptance/reject of submissions February 28, 2003: deadline for final versions of papers (including the introduction/tutorial by the session organizer) How to submit a proposal to organize a special session ? -------------------------------------------------------- Simply send an e-mail to the conference secretariat with the title of the session a brief description (+- 10 lines, to be published on the ESANN Web site) your affiliation, address, phone and fax number, e-mail address a description of your expertise in the field of the session (or a link to your personal Web page where this information can be found) Proposals are expected until August 1, 2002. Special sessions to be organized during the ESANN'2003 conference will be selected around August 15, 2002. Please mention in your e-mail if we will be able to contact you between August 1 and August 15, 2002. ======================================================== ESANN - European Symposium on Artificial Neural Networks http://www.dice.ucl.ac.be/esann * For submissions of papers, reviews,... Michel Verleysen Univ. Cath. de Louvain - Microelectronics Laboratory 3, pl. du Levant - B-1348 Louvain-la-Neuve - Belgium tel: +32 10 47 25 51 - fax: + 32 10 47 25 98 mailto:esann at dice.ucl.ac.be * Conference secretariat d-side conference services 24 av. L. Mommaerts - B-1140 Evere - Belgium tel: + 32 2 730 06 11 - fax: + 32 2 730 06 00 mailto:esann at dice.ucl.ac.be ======================================================== From M.Herbster at cs.ucl.ac.uk Tue Jul 2 17:14:33 2002 From: M.Herbster at cs.ucl.ac.uk (Mark HERBSTER) Date: Tue, 02 Jul 2002 22:14:33 +0100 Subject: Lectureship in Intelligent Systems at University College London Message-ID: <21036.1025644473@cs.ucl.ac.uk> UNIVERSITY COLLEGE LONDON Department of Computer Science Lectureship in Intelligent Systems We are seeking talented researchers who are keen to undertake innovative research embracing the application and theory of Intelligent Systems. The successful candidate is likely to already be working in Intelligent Systems or Machine Learning either in a leading university research centre or company. He/she will be expected to play a leading role in the research and teaching of Machine Learning in the Department, and especially in the development of the new MSc programme in Intelligent Systems. Applicants will have a higher degree in a relevant discipline and a strong Computer Science background. Experience of applying Machine Learning to business or industrial applications would be an advantage. The appointment will be on the Lecturer A or Lecturer B scale (22,604 to 34,671 pa including London Allowance) or on the Senior Lecturer scale (36,292 to 40,737 including London Allowance) according to experience. Further information regarding this post and the application form can be found on our website (www.cs.ucl.ac.uk/vacancies). Informal enquiries maybe directed to Professor Steve Wilbur (s.wilbur at cs.ucl.ac.uk). The closing date for applications is Tuesday, 30th July 2002. --- Mark Herbster Department of Computer Science University College London Gower Street London WC1E 6BT United Kingdom Phone: +44 (0)20 7679 3684 Fax: +44 (0)20 7387 1397 From qian at brahms.cpmc.columbia.edu Tue Jul 9 12:16:21 2002 From: qian at brahms.cpmc.columbia.edu (Ning Qian) Date: Tue, 9 Jul 2002 12:16:21 -0400 Subject: paper: structure from motion Message-ID: <200207091616.g69GGLA09177@brahms.cpmc.columbia.edu> Dear colleagues, The following paper (1.3MB) is available at: http://brahms.cpmc.columbia.edu/publications/sfm.pdf (It was published a couple of months ago but I never announced it before.) Computing relief structure from motion with a distributed velocity and disparity representation Julian Martin Fernandez, Brendon Watson, and Ning Qian, Vision Research, 2002, 42:883-898. Recent psychophysical experiments suggest that humans can only recover relief structure from motion (SFM); i.e., an object's 3D shape can only be determined up to a stretching transformation along the line of sight. Here we propose a physiologically plausible model for the computation of relief SFM, which is also applicable to the related problem of motion parallax. We assume that the perception of depth from motion is related to the firing of a subset of MT neurons tuned to both velocity and disparity. The model MT neurons are connected to each other laterally to form modulatory interactions. The overall connectivity is such that when a zero-disparity velocity pattern is fed into the system, the most responsive neurons are not those tuned to zero disparity, but instead are those having preferred disparities consistent with the relief structure of the velocity pattern. The model computes the correct relief structure under a wide range of parameters and can also reproduce the SFM illusions involving coaxial cylinders. It is consistent with the psychophysical observation that subjects with stereo impairment are also deficient in perceiving motion parallax, and with the physiological data that the responses of direction- and disparity-tuned MT cells covary with the perceived surface order of bistable SFM stimuli. Many earlier papers at http://brahms.cpmc.columbia.edu/ are now also available in PDF format. Best regards, Ning ------------------------------------------------------------- http://brahms.cpmc.columbia.edu Ning Qian, Ph. D. qian at brahms.cpmc.columbia.edu Associate Professor nq6 at columbia.edu Ctr. Neurobiology & Behavior Columbia University / NYSPI 212-543-5213 (Office) Kolb Annex Rm 730 212-543-5161 (Lab/fax) 722 W. 168th Street 212-543-6048 (Lab) New York, NY 10032, USA ------------------------------------------------------------- From jose at psychology.rutgers.edu Tue Jul 9 01:53:09 2002 From: jose at psychology.rutgers.edu (stephen j hanson) Date: 09 Jul 2002 01:53:09 -0400 Subject: Systems Admin/ Research Staff Position --Cognitive Science/Cognitive Neuroscience/Computation Message-ID: <1026194002.1980.47.camel@vaio> Systems Adminstration/UNIX/LINUX--Research Staff IMMEDIATE OPENING-- 08/01/02 PSYCHOLOGY DEPARTMENT-RUTGERS UNIVERSITY--Newark Campus-- We are searching for an individual who can adminster the computing resources of the Psychology Department at Rutgers (Newark) Resources include a network of Sun/LINUX workstations, PCs and Macs, printers, pc-voice mail system and various devices (scanners, projecters etc..). The individual will be responsible for installing and debugging software, and various routine system administration activites. About half their time will be spent in research involving Cognitive Neuroscience/Cognitive Science especially related to Connectionist networks (or Neural Networks and Computational Neuroscience. Familiarity with C programming, UNIX system internals (BSD, System V, Solaris, Linux) and Windows (XX, NT) and local area networks running TCP/IP is required. Image processing or graphics programing experience are pluses. Candidates should possess either a BS/MS in Computer Science, Cognitive Science, AI or other relevant fields or equivalent experience. We are located 15-20 minutes from Downtown Manhattan and are within minutes of the NJPAC in the heart of Newark. Salary is competitive and will be dependent upon qualifications and experience. Rutgers University is an equal opportunity affirmative action employer. Please send resumes and references to Stephen J. Hanson Department of Psychology RUMBA Labs 101 Warren Street Rutgers University Newark, New Jersey, 07102 Direct email inquiries or resumes to: jose at psychology.rutgers.edu Please indicate on SUBJECT Line: SYS ADM as a keyword. From cowen at mail.nsma.arizona.edu Thu Jul 11 20:34:50 2002 From: cowen at mail.nsma.arizona.edu (cowen@mail.nsma.arizona.edu) Date: Thu, 11 Jul 2002 17:34:50 -0700 (MST) Subject: Job Posting, Message-ID: <1026434090.3d2e242ab0db4@email.arl.arizona.edu> Job No. 24458 Appointed Position Job Title: Associate Research Scientist Department: Neural Systems Memory and Aging Division of Arizona Research Laboratories Salary: DOE Benefits Hours: 40 per week Opening: 7/10/02 Closing: 7/17/02 Position Summary: The Arizona Research Laboratories Division of Neural Systems, Memory and Aging seek candidates for the position of Associate Research Scientist. The successful candidate will be part of a collaborative research project between the University of Arizona and University of California at Davis. The primary worksite will be at the California Regional Primate Research Center in Davis, CA. This is an extended temporary position of approximately six months or more duration. This individual will be responsible for primate electrophysiology experiments using an advanced parallel recording systems developed here in Tucson (the Neuralynx Cheetah system). Minimum Qualifications: Ph.D. in Neuroscience, Behavioral Psychology or a related field. Preferred Qualifications: * Experience in neuroscience, with an emphasis in neural systems. * Familiarity with the Neuralynx Cheetah Data Acquisition System. * Experience in computer data acquisition and analysis, including Matlab. To apply, please submit a cover letter, resume and the names and contact information for three references to: Search Committee c/o Luann Snyder ARL Division of Neural Systems, Memory and Aging P.O. Box 245115 The University of Arizona Tucson, AZ 85724-5115 Please reference job number 24458. For consideration, complete requested documentation must be received by midnight of the closing date. The University of Arizona is an EEO/AA Employer-M/W/D/V From m.niranjan at dcs.shef.ac.uk Fri Jul 12 09:24:01 2002 From: m.niranjan at dcs.shef.ac.uk (Mahesan Niranjan) Date: Fri, 12 Jul 2002 14:24:01 +0100 (BST) Subject: Job Job Job Message-ID: The University of Sheffield has three Faculty openings (Lecturer / Senior Lecturer) in Computer Science. One of these is in "any area of CS" and two in "bioinformatics". Both definitions are sufficiently broad to be of interest to subscribers of this forum. In the Bioinformatics area, we have excellent collaborations with local developmental biologists and medics. http://www.shef.ac.uk/jobs/adcjobs Academic positions in the UK offer excellent freedom to pursue ideas that interest you. Other features of the job include low pay, poor student motivation, vague concept of tenure, periodic assessment of teaching and research quality in departments... If you know someone who might be interested, please point them in my direction. thanx niranjan ____________________________________________________________________ Mahesan Niranjan Phone: 44 114 222 1805 Professor of Computer Science FaX: 44 114 222 1810 University of Sheffield Email: M.Niranjan at dcs.shef.ac.uk http://www.dcs.shef.ac.uk/~niranjan ____________________________________________________________________ From marcus.hutter at gmx.net Sat Jul 13 05:39:18 2002 From: marcus.hutter at gmx.net (Marcus Hutter) Date: Sat, 13 Jul 2002 11:39:18 +0200 Subject: PhD & Postdoc Position Available Message-ID: <010901c22a51$8036fc80$78f536cb@ho> PhD & Postdoc Position Available ----------------------------------------- IDSIA, Switzerland, is seeking for one outstanding PhD student and one PostDoc with excellent mathematical skills interested in reinforcement learning, algorithmic information theory, Kolmogorov complexity, Minimal Description Length, computational complexity theory, information theory and statistics, universal Solomonoff induction, universal Levin search, sequential decision theory, adaptive control theory, and/or related areas. Possible backgrounds are computer science, physics, mathematics, etc. The initial appointment will be for 2 years. Normally there will be a prolongation. The new PhD student/PostDoc will interact with Marcus Hutter and Juergen Schmidhuber and other people at IDSIA. See http://www.idsia.ch/~marcus/idsia/phdpos1.htm for more information on the PhD position and http://www.idsia.ch/~marcus/idsia/postdoc1.htm for more information on the PostDoc position. Applicants should submit: (i) Detailed curriculum vitae, (ii) List of three references and their email addresses, (iii) Concise statement of their research interests (two pages max). Please send all documents to: Marcus Hutter, IDSIA, Galleria 2, 6928 Manno (Lugano), Switzerland. Applications can also be submitted by email to marcus at idsia.ch (2MB max). WWW pointers to ps/pdf/doc/html files are welcome. Use Firstname.Lastname.DocDescription.DocType for filename convention. Thanks for your interest Marcus Hutter, senior researcher, IDSIA Istituto Dalle Molle di Studi sull'Intelligenza Artificiale Galleria 2 CH-6928 Manno(Lugano) - Switzerland Phone: +41-91-6108668 Fax: +41-91-6108661 E-mail marcus at idsia.ch http://www.idsia.ch/~marcus --------------------------------------------------------------------------- ABOUT IDSIA. Our research focuses on artificial neural nets, reinforcement learning, complexity and generalization issues, unsupervised learning and information theory, forecasting, artificial ants, combinatorial optimization, evolutionary computation. IDSIA is small but visible, competitive, and influential. IDSIA's algorithms hold the world records for several important operations research benchmarks (see Nature 406(6791):39-42 for an overview of artificial ant algorithms developed at IDSIA). In the "X-Lab Survey" by Business Week magazine, IDSIA was ranked in fourth place in the category "COMPUTER SCIENCE - BIOLOGICALLY INSPIRED" - after the Santa Fe Institute, Stanford University, and EPFL (also in Switzerland). Its comparatively tiny size notwithstanding, IDSIA also ranked among the top ten labs worldwide in the broader category "ARTIFICIAL INTELLIGENCE". IDSIA is located near the beautiful city of Lugano in Ticino (pictures), the scenic southernmost province of Switzerland, origin of special relativity and the WWW. Milano, Italy's center of fashion and finance, is 1 hour away, Venice 3 hours. Our collaborators at CSCS (the Swiss supercomputing center) are right beneath us; we are also affiliated with the University of Lugano and SUPSI. Switzerland boasts the highest citation impact factor, the highest supercomputing capacity pc (per capita), the most Nobel prizes pc (450% of the US value), and perhaps the best chocolate. From jose at psychology.rutgers.edu Tue Jul 16 08:32:24 2002 From: jose at psychology.rutgers.edu (Stephen J. Hanson) Date: 16 Jul 2002 08:32:24 -0400 Subject: POSTDOC POSITION--IMMEDIATE OPENING Message-ID: <1026822750.3327.140.camel@madison> COGNITIVE/COMPUTATIONAL NEUROSCIENCE POSTDOCTORAL POSITION at RUTGERS UNIVERSITY, Newark Campus. The Rutgers University Mind/Brain Analysis (RUMBA) Project anticipates making one postdoctoral appointment, which is to begin in the FALL (August/September) of 2002. This positions are for a minimum of 2 years, with the possibility of continuation for 1 more year and will be in the areas of specialization of cognitive neuroscience with emphasis on the development of new paradigms and methods in neuroimaging, mathematical modeling, signal processing or data analysis in functional brain imaging. Particular interest is in methods and algorithms for fusion of EEG/fMRI. Applications are welcomed begining immediately and review will continue until the position is filled. Rutgers University is an equal opportunity/affirmative action employer. Qualified women and minority candidates are especially encouraged to apply. Send CV and three letters of recommendation and 1 reprint to Professor S.J. Hanson, Department of Psychology, Rutgers University, Newark, NJ 07102. Email enquiry can be made to jose at psychology.rutgers.edu please put "RUMBA POSTDOC" in your subject field also see http://www.rumba.rutgers.edu. From Zoubin at gatsby.ucl.ac.uk Wed Jul 17 07:44:58 2002 From: Zoubin at gatsby.ucl.ac.uk (Zoubin Ghahramani) Date: Wed, 17 Jul 2002 12:44:58 +0100 (BST) Subject: Faculty Position in Intelligent Systems / Machine Learning in London Message-ID: <200207171144.MAA07760@cajal.gatsby.ucl.ac.uk> For those of you who may not have seen this ad for a faculty position in Intelligent Systems in the Department of Computer Science at University College London (UCL), I've enclosed it below. I'd like to add that the MSc in Intelligent Systems was established and is taught jointly by the Computer Science department and the Gatsby Unit at UCL (see http://www.gatsby.ucl.ac.uk). We hope that the successful applicant for this post will have strong interactions with members of the Gatsby Unit, adding to and complementing our key strengths in learning theory, Bayesian methods, reinforcement learning, pattern recognition, kernel machines and graphical models. -Zoubin Ghahramani University College London ---------------------------------------------------------------------- UNIVERSITY COLLEGE LONDON Department of Computer Science We are seeking talented researchers who are keen to undertake innovative research embracing the application and theory of Intelligent Systems. The successful candidate is likely to already be working in Intelligent Systems either in a leading university research centre or company. He/she will be expected to play a leading role in the research and teaching of Intelligent Systems in the Department, and especially in the development of the new MSc programme in Intelligent Systems. Applicants will have a higher degree in a relevant discipline and a strong Computer Science background. Experience of applying Intelligent Systems to business or industrial applications would be an advantage. The appointment will be on the Lecturer A or Lecturer B scale (?22,604 to ?34,671 pa including London Allowance) according to experience. Further information regarding this post and the application form can be found on our website (http://www.cs.ucl.ac.uk/vacancies). Informal enquiries may be directed to Professor Steve Wilbur (s.wilbur at cs.ucl.acuk). The closing date for applications is Tuesday, 30th July 2002. Taking Action for Equality Claire Cawley Human Resources Administrator and PA to Professor Wilbur, Head of Department Department of Computer Science University College London Gower Street London WC1E 6BT T: 020 7679 3676 (internal: 33676) F: 020 7387 1397 E: c.cawley at cs.ucl.ac.uk I: www.cs.ucl.ac.uk/staff/C.Cawley/ From niki at cis.ohio-state.edu Wed Jul 17 14:38:51 2002 From: niki at cis.ohio-state.edu (Nicole Roman) Date: Wed, 17 Jul 2002 14:38:51 -0400 Subject: Tech report on location-based segregation Message-ID: <3D35B9BB.23CB64BD@cis.ohio-state.edu> Dear Colleagues, It is my pleasure to announce the availability of the following technical report. Thanks for your attention, Nicoleta Roman ************************************ "Speech segregation based on sound localization", Technical Report #16, June 2002. Department of Computer and Information Science The Ohio State University Nicoleta Roman, The Ohio State University DeLiang Wang, The Ohio State University Guy J. Brown, University of Sheffield ************************************* Abstract --------- At a cocktail party, we can selectively attend to a single voice and filter out all the other acoustical interferences. How to simulate this perceptual ability remains a great challenge. This paper describes a novel machine learning approach to speech segregation, in which a target speech signal is separated from interfering sounds using spatial location cues: interaural time differences (ITD) and interaural intensity differences (IID). The auditory masking effect motivates the notion of an ?ideal? time-frequency binary mask, which selects the target if it is stronger than the interference in a local time-frequency (T-F) unit. We observe that within a narrow frequency band, modifications to the relative strength of the target source with respect to the interference trigger systematic deviations for ITD and IID. For a given spatial configuration, this interaction produces characteristic clustering in the binaural feature space. Consequently, we perform pattern classification in order to estimate ideal binary masks. A systematic evaluation shows that the resulting system produces masks very close to ideal binary ones, and gives a significant improvement in performance over an existing approach, as quantified by changes in signal-to-noise ratio before and after segregation. ************************************** The manuscript is available for download at: ftp://ftp.cis.ohio-state.edu/pub/tech-report/2002/TR16.pdf Related sound demos can be found at: http://www.cis.ohio-state.edu/~niki/soundemo.html A preliminary version of this work is included in the Proceedings of 2002 ICASSP. From arbib at pollux.usc.edu Wed Jul 17 21:53:27 2002 From: arbib at pollux.usc.edu (Michael Arbib) Date: Thu, 18 Jul 2002 09:53:27 +0800 Subject: Post-Doctoral Traineeship in Computational and Cognitive Neuroscience at the University of Southern California Message-ID: <200207180155.JAA06657@bilby.cs.uwa.edu.au> The Neuroscience Program at the University of Southern California (http://www.usc.edu/dept/nbio/ngp/) has an active program of research in a broad range of areas of Computational and Cognitive Neuroscience including vision, motor control, visuomotor coordination, linguistics (with an active new focus on an action-oriented approach to the evolution of brain mechanisms of language), neuroinformatics, neural engineering, and memory and learning. The Training Grant in Computational and Cognitive Neuroscience at USC has one opening for a postdoctoral trainee, effective immediately. Candidates must be US nationals or permanent residents. In addition to the NIH-mandated salary, the trainee will receive 4 units/year tuition, payment of mandatory student fees, a travel allowance of $500 and $3500 for training related expenses (M&S). Candidates with a strong interest in Computational and/or Cognitive Neuroscience should email an application to Michael Arbib (arbib at pollux.usc.edu) with the following materials: 1. Name, email, and statement of US citizenship or permanent resident status 2. Undergraduate institution, major, GPA and GRE. 3. Ph.D. thesis title and abstract, institution and years of study, list of all graduate courses taken (with grades). 4. A one page statement of interest in Computational and/or Cognitive Neuroscience as an area for postdoctoral research, with a list of one or more possible supervisors at USC with reasons for your choice(s). 5. A list of publications, presentations, and awards. In addition, the applicant should ask three researchers with relevant expertise to send Dr. Arbib a half-page email attesting to the applicant's suitability for a traineeship. Applications and support letters must be received by August 19, 2002 for consideration for Traineeships for 12 months beginning either September 1, 2002 or January 1, 2003 (candidates should indicate for which of these periods they are available). ******************************************************* Michael A. Arbib USC Brain Project University of Southern California Los Angeles, CA 90089-2520 Phone (213) 740-9220 FAX (213) 740-5687 arbib at pollux.usc.edu [Additional details for Express Mail: Hedco Neuroscience Building, Room 5, 3614 Watt Way; Phone Contact 213-740-1176.] ______________________ http://www-hbp.usc.edu/people/arbib.htm http://www-hbp.usc.edu/ http://mitpress.mit.edu/e-books/HBTNN2e http://www.cs.usc.edu/ ******************************************************* From d.mareschal at bbk.ac.uk Wed Jul 17 12:52:01 2002 From: d.mareschal at bbk.ac.uk (Denis Mareschal) Date: Wed, 17 Jul 2002 17:52:01 +0100 Subject: 3 YEAR PHD POSITION in Belgium Message-ID: Please forward this to any interested parties. Please DO NOT REPLY DIRECTLY TO ME. cheers, Denis ============================ 3 year PHD Position A 3-year position for a graduate student in cognitive science, funded by a research grant from the European Commission, is available immediately at the Free University of Brussels (ULB)in the laboratory of Axel Cleeremans. Dr. Cleeremans' lab is primarily in the area of implicit learning (i.e., learning without being consciously aware of learning having taken place), adn consciousness. The research carried out by his group is cutting edge and has been published in some of the foremost international journals in the world (e.g., Nature Neuroscience). The research in this lab uses a combination of neural network modelling and experimental testing to better understand observable behavioral phenomena. In the area of connectionist modeling of psychological processes, Cleeremans' lab is one of the best known in Europe. It is comprised of a relatively small and tight-knit group of researchers, and is well funded and well equipped. Details of the research activities in this lab can be found at : http://srsc.ulb.ac.be/axcWWW/axc.html The European Commission funding for a qualified graduate student will run until February 2005. The candidate will be expected to enroll in the Ph.D. program at ULB, under Dr. Cleermans' supervision. A background in experimental psychology and a good grounding in computer programming (e.g., C++, JAVA, Delphi, or similar languages) are the only pre-requisites for this position. Funding is available as of September 1st. Applications will continue to be taken an appropriate candidate is found. Some funding restrictions apply. The candidate must be less than 35 years old, must be a citizen (or long temr resident) of one of the member or associate-member states of the European Union. However, cannot be a Belgian citizen and cannot have resided in Belgium for more than 1 year in the last two years. For further information, please contact: - Dr. Axel Cleeremans directly: axcleer at ulb.ac.be, - Dr. Robert French, at the University of Liege in Belgium, director of the EC project: rfrench at ulg.ac.be For more information on the EU research project, and the partners in the project, please consult the following URL's: http://www.ulg.ac.be/cogsci/bmlf/ and for a detailed description of the Project itself: http://www.ulg.ac.be/rfrench/bmlf.pdf ================================================= Dr. Denis Mareschal Centre for Brain and Cognitive Development School of Psychology Birkbeck College University of London Malet St., London WC1E 7HX, UK tel +44 020 7631-6582/6207 fax +44 020 7631-6312 http://www.psyc.bbk.ac.uk/staff/dm.html ================================================= From juergen at idsia.ch Thu Jul 18 10:18:20 2002 From: juergen at idsia.ch (Juergen Schmidhuber) Date: Thu, 18 Jul 2002 16:18:20 +0200 Subject: near-optimal computable predictions Message-ID: <3D36CE2C.E3CD6023@idsia.ch> The Speed Prior: a new simplicity measure yielding near-optimal computable predictions (Juergen Schmidhuber, IDSIA) In J. Kivinen and R. H. Sloan, eds, Proc. 15th Annual Conf. on Computational Learning Theory (COLT), 216-228, Springer, 2002; based on section 6 of http://arXiv.org/abs/quant-ph/0011122 (2000) http://www.idsia.ch/~juergen/speedprior.html ftp://ftp.idsia.ch/pub/juergen/colt.ps Solomonoff's optimal but noncomputable method for inductive inference assumes that observation sequences x are drawn from an recursive prior distribution mu(x). Instead of using the unknown mu(x) he predicts using the celebrated universal enumerable prior M(x) which for all x exceeds any recursive mu(x), save for a constant factor independent of x. The simplicity measure M(x) naturally implements "Occam's razor" and is closely related to the Kolmogorov complexity of x. However, M assigns high probability to certain data x that are extremely hard to compute. This does not match our intuitive notion of simplicity. Here we suggest a more plausible measure derived from the fastest way of computing data. In absence of contrarian evidence, we assume that the physical world is generated by a computational process, and that any possibly infinite sequence of observations is therefore computable in the limit (this assumption is more radical and stronger than Solomonoff's). Then we replace M by the novel Speed Prior S, under which the cumulative a priori probability of all data whose computation through an optimal algorithm requires more than O(n) resources is 1/n. We show that the Speed Prior allows for deriving a computable strategy for optimal prediction of future y, given past x. Then we consider the case that the data actually stem from a nonoptimal, unknown computational process, and use Hutter's recent results to derive excellent expected loss bounds for S-based inductive inference. Assuming our own universe is sampled from S, we predict: it won't get many times older than it is now; large scale quantum computation won't work well; beta decay is not random but due to some fast pseudo-random generator which we should try to discover. Juergen Schmidhuber http://www.idsia.ch/~juergen From levy at cs.brandeis.edu Thu Jul 18 06:10:37 2002 From: levy at cs.brandeis.edu (Simon Levy) Date: Thu, 18 Jul 2002 06:10:37 -0400 Subject: Ph.D. thesis announcement: Infinite RAAM Message-ID: <3D36941D.2090707@cs.brandeis.edu> Levy, Simon D. (2002). Infinite RAAM: Initial Investigations into a Fractal Basis for Cognition. Ph.D. Thesis, Brandeis University, July 2002. Abstract This thesis attempts to provide an answer to the question ``What is the mathematical basis of cognitive representations?'' The answer we present is a novel connectionist framework called Infinite RAAM. We show how this framework satisfies the cognitive requirements of systematicity, compositionality, and scalable representational capacity, while also exhibiting ``natural'' properties like learnability, generalization, and inductive bias. The contributions of this work are twofold: First, Infinite RAAM shows how connectionist models can exhibit infinite competence for interesting cognitive domains like language. Second, our attractor-based learning algorithm provides a way of learning structured cognitive representations, with robust decoding and generalization. Both results come from allowing the dynamics of the network to devise emergent representations during learning. An appendix provides Matlab code for the experiments described in the thesis. Keywords: Neural Networks, Fractals, Connectionism, Language, Grammar. Postscript: http://www.demo.cs.brandeis.edu/papers/levythesis.ps Gzipped: http://www.demo.cs.brandeis.edu/papers/levythesis.ps.gz PDF: http://www.demo.cs.brandeis.edu/papers/levythesis.pdf From ckiw at dai.ed.ac.uk Fri Jul 19 09:24:54 2002 From: ckiw at dai.ed.ac.uk (Chris Williams) Date: Fri, 19 Jul 2002 14:24:54 +0100 (BST) Subject: Faculty Positions at the University of Edinburgh Message-ID: The School of Informatics invites applications from candidates of international standing for three appointments at the level of either Reader or Lecturer. Outstanding candidates in all the following areas are invited to apply: Algorithms; Cognitive Systems; Computer Vision, Graphics and Robotics; Computing Systems Architecture; Knowledge Representation and Reasoning; Machine Learning and Probabilistic Modelling; and Software Engineering. [The Lecturer and Reader posts are roughly equivalent to US Assistant and Associate Professor levels.] We invite applicants from all areas of machine learning and probabilistic modelling. We have particular interest in probabilistic graphical models, learning applied to computer vision, and data mining of complex data types. Research in machine learning and probabilistic modelling occurs in many areas of the School, with the Institute for Adaptive and Neural Computation acting as a hub for these activities. These research activities are supported by collaborations within and outwith the university, for example with the Department of Child Life and Health (condition monitoring on premature babies) and with the Royal Observatory of Edinburgh (astronomical data mining). Informal enquiries with regard to Machine Learning and Probabilistic Modelling may be made to Dr Chris Williams (c.k.i.williams at ed.ac.uk). Informatics at Edinburgh is one of the top-ranked departments in the United Kingdom. Candidates should demonstrate a world-class research record and both interest and ability in teaching. Candidates for a readership will be expected to demonstrate the ability to take on research leadership in their respective area. Informal enquiries to Bonnie Webber, +44 131 650 4190 (bonnie.webber at ed.ac.uk) or to Michael Fourman, +44 131 650 2703 (hod at informatics.ed.ac.uk). Further information can be found at both http://www.informatics.ed.ac.uk/events/vacancies/ and http://www.jobs.ed.ac.uk/. Salary scale: Lecturer 20,470 - 32,537 pounds p.a. (under review) Reader 34,158 - 38,603 pounds p.a. (under review) Please quote Ref: 311616 Letters of application should include a curriculum vitae and the names and addresses of 3 referees. Please include fax numbers and email addresses for referees if possible. Applications should be addressed to Division of Informatics (c/o Ms. Eleanor Kerse), University of Edinburgh, and sent, to arrive not later than 23 August 2002, by post (80 South Bridge, Edinburgh EH1 1HN, UK), fax (+44 (0)131 650 6516), or email (hod at informatics.ed.ac.uk). Applications can be made on-line through http://www.jobs.ed.ac.uk/. Closing date: 23 August 2002 Dr Chris Williams ckiw at dai.ed.ac.uk Institute for Adaptive and Neural Computation Division of Informatics, University of Edinburgh 5 Forrest Hill, Edinburgh EH1 2QL, Scotland, UK fax: +44 131 650 6899 tel: (direct) +44 131 651 1212 (department switchboard) +44 131 650 3090 http://www.dai.ed.ac.uk/homes/ckiw/ From mm at santafe.edu Mon Jul 22 15:18:43 2002 From: mm at santafe.edu (Melanie Mitchell) Date: Mon, 22 Jul 2002 13:18:43 -0600 (MDT) Subject: Graduate Research Positions Message-ID: <200207221918.g6MJIhv08743@taos.santafe.edu> I have openings for two graduate research assistants to work on a computer model of analogy-making between visual images. This work will build on the "Copycat" model of Hofstadter and Mitchell and will incorporate other approaches to high-level perception and image understanding, including those inspired by the field of "complex adaptive systems". More information about the project can be found at http://www.santafe.edu/~mm/analogy-vision.html. A recent paper describing the Copycat model, "Analogy-making as a complex adaptive system", can be downloaded from http://www.santafe.edu/~mm/paper-abstracts.html#amcas. Applicants must be willing to pursue a graduate degree in Computer Science and Engineering at the OGI School of Science and Engineering, Oregon Health & Science University, near Portland, Oregon, where I will be joining the faculty. The department web pages can be found at http://www.cse.ogi.edu. Proficiency in C, C++, or another high-level programming language is required. Background in cognitive science, psychology, computer science, mathematics, image processing and computer vision, and/or biology would be helpful. The assistanceship will cover tuition and stipend. To apply, send a resume with your research interests, list of relevant course work or experience, programming experience and languages, and any other information you think would be relevant, and the names and contact information of at least two professors or scientists who will act as references. Please send this information in electronic form to mm at santafe.edu. Applications will be considered until the positions are filled. Students of any nationality may apply. OGI is an equal opportunity employer and particularly welcomes applications from women and minority candidates. ----------------------------------- Melanie Mitchell Associate Professor Department of Computer Science and Engineering OGI School of Science & Engineering Oregon Health & Science University 20000 NW Walker Road Beaverton, OR 97006 E-mail: mm at santafe.edu From hadley at cs.sfu.ca Thu Jul 25 15:48:47 2002 From: hadley at cs.sfu.ca (Bob Hadley) Date: Thu, 25 Jul 2002 12:48:47 -0700 (PDT) Subject: Systematicity & Fallacies: Boden & Niklasson Message-ID: <200207251948.g6PJmll17978@css.css.sfu.ca> The Fallacy of Equivocation: Boden and Niklasson. In a fairly recent paper (Connection Science, Vol. 12, 2000), Boden and Niklasson purport to demonstrate that a collection of connectionist networks (call them c-nets) can display an important type of Strong Semantic Systematicity. They make frequent references to my 1994 definitions of semantic systematicity and to my papers on this important topic. They also acknowledge that in 1994 I published definite reservations about claims by Niklasson and van Gelder to have produced a connectionist system that displays strong systematicity. In their recent (2000) paper, Boden and Niklasson purport to have answered my reservations by producing a case where a "novel test sentence" is assigned an appropriate meaning representation by previously trained c-nets. Readers may recall that my 1994 definition of strong semantic systematicity required that the "previously trained c-net" must assign an appropriate (and correct) meaning representation to a novel test sentence which contains PREVIOUSLY KNOWN words in at least one novel position. In contrast to this requirement, the putative novel test sentence that Boden and Niklasson employ does not present any previously known words in a novel position. Rather, it presents a purportedly novel word in a known position. However, there is a much more serious problem with their "novel test sentence" (call this sentence S). Here's the problem: The supposed novel sentence S does not produce a correct response when it is first presented to the trained c-net. So, Boden and Niklasson proceed to TRAIN the c-net on the sentence S for an additional 1000 epochs (over and above the earlier training phase). In this latter training phase, only S is presented as input, and backpropagation is employed. Once this further training is complete, Boden and Niklasson contend that a "novel" word in S has now been assigned a meaning representation which they believe to be correct. But, of course, S is no longer a "novel test sentence" at this stage. The c-net has been subjected to intensive training upon S, and only after this further training is complete are Boden and Niklasson able to claim success. Given this, for Boden and Niklasson to describe S as a novel test sentence is (to express the matter diplomatically) to committ a serious instance of the fallacy of equivocation. Indeed, I find it difficult to believe that Boden and Niklasson could be unaware that, as most connectionists use the phrase "test data" (or "novel test sentence"), sentence S is NOT a novel test sentence at all. For this reason, it astonishes me that Boden and Niklasson claim that they have NOW produced an experimental result that satisfactorily answers my 1994 reservations about the results published by Niklasson and van Gelder. My 1994 reservations involved my 1994 definition of strong systematicity, and that definition employed "novel test sentence" in the sense that connectionists commonly employ. At best, Boden and Niklasson are assigning some new, and surprising sense to that phrase -- hence the fallacy of equivocation. I believe there are other serious problems with Boden and Niklasson's (2000) paper, and I am presently writing a detailed critique of that paper. I'll make my new paper available on the internet within a few weeks. Look for a notice of my new critique on "Connectionist List" or send me an email request for the pdf file. In astonishment, Bob Hadley Reference: Boden, M. and Niklasson, L. (2000) "Semantic Systematicity and Context in Connectionist Networks", Connection Science, Vol. 12(2), pp. 111-142. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Robert F. Hadley (Bob) Phone: 604-291-4488 Professor email: hadley at cs.sfu.ca School of Computing Science and Cognitive Science Program Simon Fraser University Burnaby, B.C. V5A 1S6 Canada Web page: www.cs.sfu.ca/~hadley/ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ From sfr at unipg.it Thu Jul 25 06:36:40 2002 From: sfr at unipg.it (Simone G.O. Fiori (An)) Date: Thu, 25 Jul 2002 12:36:40 +0200 Subject: Mathematical analysis of unsupervised learning systems. Message-ID: <1.5.4.32.20020725103640.01bc3504@unipg.it> Dear Colleagues, I would like to announce the availability of three new papers devoted to the mathematical analysis of unsupervised neural learning systems. Sincerely, Simone Fiori Unsupervised Neural Learning on Lie Group ============================================================================= Author: S. Fiori, Faculty of Engineering, University of Perugia (Italy) Journal: International Journal of Neural Systems Abstract: The present paper aims at introducing the concepts and mathematical details of unsupervised neural learning with orthonormality constrains. The neural structures considered are single non-linear layers and the learnable parameters are organized in matrices, as usual, which gives the parameters spaces the geometrical structure of the Euclidean manifold. The constraint of orthonormality for the connection-matrices further restrict the parameters spaces to differential manifolds such as the orthogonal group, the compact Stiefel manifold and its extensions. For these reasons, the instruments for characterizing and studying the behavior of learning equations for these particular networks are provided by the differential geometry of Lie groups. Although the considered class of learning theories is very general, in the present paper special attention is paid to unsupervised learning paradigms. Download from: http://www.unipg.it/~sfr/publications/IJNS02.ps [46 pages, 565 KB] Notes on Bell-Sejnowski PDF-Matching Neuron ============================================================================= Author: S. Fiori, Faculty of Engineering, University of Perugia (Italy) Journal: Neural Computation Extended abstract: Independent component analysis (ICA) is an emerging neural signal processing technique that allows representing sets of signals as linear combinations of statistically independent bases. In particular, the results of recent investigations about the statistical properties of natural images in relation to the properties of simple cells in V1, suggest that these cells learn to form spatial filters that perform an independent component analysis of the images. One of the most interesting aspects of the ICA theory proposed by Bell and Sejnowski (1996) is the emerging ability of the neurons in the structural model to align to the statistical distributions of the stimuli. Such observation was successfully exploited in order to design different learning rules for blind separation, blind deconvolution and probability density function estimation. In the present paper we consider the basic Bell-Sejnowski class of neuron models and recall the maximum-entropy adapting formulas. By properly selecting a model in the class, that gives rise to tractable mathematics, we are able to present the closed-form expressions of the learning equations, that we particularize for some special excitations. Our main goal is to discuss the features of the neuron-model in an analytical way, in order to gain a deeper insight into the behavior of the equations governing information-theoretic non-linear unit learning. Download from: http://www.unipg.it/~sfr/publications/NeCo2002.zip [8 pages, 576 KB] Information-Theoretic Learning for FAN Network Applied to Eterokurtic Component Analysis ============================================================================= Author: S. Fiori, Faculty of Engineering, University of Perugia (Italy) Journal: IEE Proceedings -- Image, Vision, and Signal Processing Extended abstract: In this paper we deal with instantaneous linear mixtures and focus on the stream of INFOMAX learning algorithms. The paper is devoted to the separation of mixed independent signals from their linear mixtures when the observations are mixed plati-kurtic and lepto-kurtic signals, that is referred to as hybrid or eterokurtic sources problem. We propose the use of networks formed by unsupervised adaptive activation function neurons (FAN), which provide a natural way of estimating the high-order statistical features required to achieve separation. Through numerical and analytical studies the effectiveness of the presented approach is also illustrated and discussed. In Section 2 the problem at hand is formally presented and the adaptive activation function structure is shown to emerge as a natural solution. In Section 3 the general unsupervised learning theory for the FAN neuron is derived, along with a closely-related one, based on a mixture-of-kernel architecture, which is considered for further numerical and architectural comparisons. In Section 4 four different FAN structures are proposed and discussed, while Section 5 is devoted to computer simulations and comparisons. Download from: http://www.unipg.it/~sfr/publications/EKA2002.zip [27 pages, 483 KB] =================================================== Dr Simone Fiori (Mr, EE, PhD)- Assistant Professor Faculty of Engineering - Perugia University Via Pentima bassa, 21 - 05100 TERNI (Italy) eMail: sfr at unipg.it - Fax: +39 0744 492925 Web: http://www.unipg.it/~sfr/ =================================================== From J.A.Bullinaria at cs.bham.ac.uk Fri Jul 26 08:02:47 2002 From: J.A.Bullinaria at cs.bham.ac.uk (John A Bullinaria) Date: Fri, 26 Jul 2002 13:02:47 +0100 (BST) Subject: PhD Studentship/Scholarship Message-ID: =========================================================================== Preliminary Announcement of an Anticipated PhD Studentship/Scholarship ---------------------------------------------------------------------- The School of Computer Science, the University of Birmingham, UK, anticipates (subject to final approval and exchange of contract) setting up a highly competitive PhD Studentship/Scholarship in the Natural Computation group to work on the project, "Automatic Problem Decomposition Using Co-evolution and Ensembles", funded by Honda R&D Europe (Deutschland) GmbH. The PhD Studentship/Scholarship is valued at approximately 17,000 pounds to 20,000 pounds per annum (covering tuition fees, maintenance costs, and travel to the Honda R&D Europe office in Germany) for up to three years (subject to the satisfactory progress of the postholder). In addition, conference travels to present accepted papers may also be funded by Honda R&D Europe or the School of Computer Science on a case by case basis. We are looking for an outstanding candidate to take up this PhD Studentship/Scholarship. The successful candidate must have a first class honours or equivalent in computer science or a closely related field. If your grades are not classified as in a British honours degree, you need to show that you are within at least the top 3% of your year in your average mark. If you have some work/research experience already, please send your best paper with your application. The successful candidate is expected to spend approximately 8 weeks each year at Honda R&D Europe. The School of Computer Science at the University of Birmingham has a strong research group in natural computation, with eight members of academic staff (six faculty and two research fellows) currently specialising in this field: Dr. John Bullinaria (Neural Networks, Evolutionary Computation, Cog.Sci.) Dr. Ke Chen (Neural Networks, Pattern Recognition, Machine Perception) Dr. Aniko Ekart (Genetic Programming, AI, Machine Learning) Dr. Jun He (Evolutionary Computation, Artificial Immune Systems) Dr. Julian Miller (Evolutionary Computation, Machine Learning) Dr. Jon Rowe (Evolutionary Computation, AI) Dr. Thorsten Schnier (Evolutionary Computation, Engineering Design) Prof. Xin Yao (Evolutionary Computation, NNs, Nature Inspired Comp.) Other staff members also working in these areas include Prof. Aaron Sloman (evolvable architectures of mind, co-evolution, interacting niches), Dr. Jeremy Wyatt (evolutionary robotics, classifier systems), and Dr Ela Claridge (evolutionary image processing). There are more than a dozen PhD students currently working in this field. For further information on the technical issues related to this PhD Studentship/Scholarship, please contact Prof. Xin Yao (x.yao at cs.bham.ac.uk). For application and anything else, please contact Dr Peter Hancox, the Research Student Admissions Tutor (p.j.hancox at cs.bham.ac.uk). More information about the PhD programme in the School of Computer Science can be found at http://www.cs.bham.ac.uk/study/postgraduate-research/. ============================================================================ From Gunnar.Raetsch at anu.edu.au Mon Jul 29 03:55:07 2002 From: Gunnar.Raetsch at anu.edu.au (Gunnar Raetsch) Date: Mon, 29 Jul 2002 17:55:07 +1000 Subject: Machine Learning Summer School 2003 Message-ID: <3D44F4DB.20305@anu.edu.au> Machine Learning Summer School The Australian National University, Canberra, Australia 2nd - 15th of February, 2003 ----------------------------------------------------------------------- We would like to inform you that *The Australian National University* will be hosting a Machine Learning Summer School. The School will consist of three courses and a series of special talks and short courses taught by experts from Australia and overseas. The School will be held between February 2 and February 15, 2003. It is suitable for all levels, both for people without previous knowledge in machine learning and those wishing to broaden their expertise in this area. It will allow the participants to get in touch with international experts in this field. The list of courses includes: + Information Geometry, *Shun-Ichi Amari*, RIKEN + Unsupervised Learning, *Zoubin Ghahramani*, Gatsby Institute + Concentration Inequalities, *Gabor Lugosi*, Pompeu Fabra University We are offering a limited number of scholarships (AU$500 and a waiver of the registration cost) for students with a strong academic background. The application deadline is December 1, 2002. Students who are interested should include their CV with their application. Moreover, students may sign up for volunteer work in order to have their registration fees waived. The registration cost of the school is AU$1,200 per person for participants from industry and AU$450 per person for academics. Students are eligible for a further discount and may register for AU$150 per person. All prices are in Australian dollars and include GST. The closing date for early registrations is, December 31, 2001. Registrations received after this date will be subject to 33% surcharge. For further information see our website at http://mlg.anu.edu.au/summer2003 or send e-mail to ml2003 at mlg.anu.edu.au. Regards, Shahar Mendelson, Gunnar Raetsch and Alex Smola -- +-----------------------------------------------------------------+ Gunnar Raetsch http://mlg.anu.edu.au/~raetsch Australian National University mailto:Gunnar.Raetsch at anu.edu.au Research School for Information Tel: (+61) 2 6125-8647 Sciences and Engineering Fax: (+61) 2 6125-8651 Canberra, ACT 0200, Australia From terry at salk.edu Mon Jul 29 18:26:17 2002 From: terry at salk.edu (Terry Sejnowski) Date: Mon, 29 Jul 2002 15:26:17 -0700 (PDT) Subject: NEURAL COMPUTATION 14:9 In-Reply-To: <200206010019.g510Jd343973@purkinje.salk.edu> Message-ID: <200207292226.g6TMQHX88979@purkinje.salk.edu> Neural Computation - Contents - Volume 14, Number 9 - September 1, 2002 ARTICLE Scalable Hybrid Computation with Spikes Rahul Sarpeshkar and Micah O'Halloran NOTES Learning Nonregular Languages: A Comparison of Simple Recurrent Networks and LSTM J. Schmidhuber, F. Gers and D. Eck Center-Crossing Recurrent Neural Networks for the Evolution of Rhythmic Behavior Boonyanit Mathayomchan, Randall D. Beer Reply to Carreira-Perpinan and Goodhill Nicholas V. Swindale, Doron Shoham, Amiram Grinvald, Tobias Bonhoeffer and Mark Hubener LETTERS Dynamics of the Firing Probability of Noisy Integrate-and-Fire Neurons Nicolas Fourcaud and Nicolas Brunel Integrate-and-Fire Neurons Driven by Correlated Stochastic Input Emilio Salinas and Terrence Sejnowski Preintegration Lateral Inhibition Enhances Unsupervised Learning M. W. Spratling and M.H. Johnson On Optimality in Auditory Information Processing Mattias F. Karlsson and John W. C. Robinson Computational Capacity of an Odorant Discriminator: The Linear Separability of Curves N. Caticha, J. E. Palo Tejada, D. Lancet and E. Domany Mixture of Experts Classification Using a Hierarchical Mixture Model Michalis K. Titsias and Aristidis Likas On the Emergence of Rules in Neural Networks Stephen Jose Hanson and Michiro Negishi ----- ON-LINE - http://neco.mitpress.org/ SUBSCRIPTIONS - 2002 - VOLUME 14 - 12 ISSUES USA Canada* Other Countries Student/Retired $60 $64.20 $108 Individual $88 $94.16 $136 Institution $506 $451.42 $554 * includes 7% GST MIT Press Journals, 5 Cambridge Center, Cambridge, MA 02142-9902. Tel: (617) 253-2889 FAX: (617) 577-1545 journals-orders at mit.edu ----- From d.mareschal at bbk.ac.uk Tue Jul 30 07:46:08 2002 From: d.mareschal at bbk.ac.uk (Denis Mareschal) Date: Tue, 30 Jul 2002 12:46:08 +0100 Subject: No subject Message-ID: Dear all, The following recent special issue of Developmental Science may be of interest to readers of this list. It provides an introduction to and overview of current cutting-edge brain imaging methods for use during human infancy. cheers, Denis ======================================================= Contents List Developmental Science 5:3 Special Issue Invited Co-Editors: BJ Casey and Michelle de Haan Title: Imaging Techniques and their Application to Developmental Science 1. Preface: Mark Johnson 2. Introduction: BJ Casey & Michelle de Haan 3. Basic Principles of MRI and Morphometry Studies of Human Brain Development: D. Kennedy, N. Makris, M. R. Herbert, T. Takahashi & V. S. Caviness Jr. 4. Magnetic resonance approaches to the identification of focal pathophysiology in children with brain disease: D. Gadian 5. Monitoring brain development with Quantitative Diffusion Tensor Imaging: A. Ulug 6. Mapping the development of white matter tracts with Diffusion Tensor Imaging: T. Li 7. Functional Magnetic Resonance Imaging : Basic principles of and application to developmental science: BJ Casey, M. Davidson & B. Rosen 8. Application of Pharmacological fMRI with developmental disorders: C. Vaidya 9. Basic principles and applications of ERP/EEG/Intracranial Methods: M. Taylor, P. Sabatier & T. Baldeweg 10. Applications of ERP and fMRI techniques to developmental science: M. de Haan & K. Thomas 11. Functional brain imaging of childhood clinical disorders with PET and SPECT: A. De Volder 12. Magnetoencephalography in pediatric neuroimaging: R. Paetau 13. Basic principles of Optical Imaging and application to the study of infant development: J. Meek 14. Transcranial Magnetic Stimulation in child psychiatry: Disturbed motor system excitability in hypermotoric syndromes: G. H. Moll 15. Commentary: A developmental psychologist looks ahead at developmental neuroimaging research: L. Spelke More details detals about Developmental Science can be obtained form http://www.blackwellpublishers.co.uk/asp/journal.asp?ref=1363-755X ================================================= Dr. Denis Mareschal Centre for Brain and Cognitive Development School of Psychology Birkbeck College University of London Malet St., London WC1E 7HX, UK tel +44 (0)20 7631-6582/6226 reception: 6207 fax +44 (0)20 7631-6312 http://www.psyc.bbk.ac.uk/staff/dm.html ================================================= From smyth at ics.uci.edu Tue Jul 30 14:06:01 2002 From: smyth at ics.uci.edu (Padhraic Smyth) Date: Tue, 30 Jul 2002 11:06:01 -0700 Subject: 2 research positions in machine learning at UC Irvine Message-ID: <3D46D589.F83F985C@ics.uci.edu> We anticipate having 2 postdoc positions in the area of probabilistic learning at UC Irvine for a new NSF-funded project, involving applications to Web and text data. See details below. I would like to encourage readers of this list to apply if interested. Please feel free to distribute to other colleagues. Apologies in advance if you receive multiple copies from cross-posting. Padhraic Smyth Postdoctoral Research Positions in Machine Learning Department of Information and Computer Science University of California, Irvine The Department of Information and Computer Science (ICS) anticipates two full-time research positions available in the area of machine learning and data mining. Responsibilities will include conducting advanced research in predictive and stochastic modeling of large data sets involving events over time (such as Web log data) as well as algorithm development for information extraction and analysis of text streams. Further responsibilities include research demonstrations and presentations, as well as collaborating with graduate students and faculty in the ICS department. Applicants must have earned a Ph.D. in Computer Science, Electrical Engineering, Applied Mathematics, Statistics, or a closely-related discipline, with an emphasis on machine learning or applied statistics. Applicants must have specific knowledge of probabilistic learning methods (such as the EM algorithm and Bayesian learning). Experience with analysis and modeling of large data sets (such as text and Web data) is desirable. The salary range for this position is $45,048 to $54,240 annually, commensurate with training and experience. The initial appointment will be for a twelve-month period with extensions for further years dependent in part on the availability of extra-mural funding. Interested applicants should respond no later than the closing data of September 30th, 2002, by forwarding a cover letter, Curriculum Vitae, and the names of three references to: Professor Padhraic Smyth Department of Information and Computer Science University of California, Irvine CA 92697-3425 or via email to smyth at ics.uci.edu The University of California, Irvine, is an Equal Opportunity Employer, committed to excellence through diversity. From janet at psy.uq.edu.au Tue Jul 30 23:55:52 2002 From: janet at psy.uq.edu.au (Janet Wiles) Date: Wed, 31 Jul 2002 13:55:52 +1000 (EST) Subject: Faculty Position at the University of Queensland Message-ID: I would like to encourage readers of this list interested in computational and cognitive neuroscience to apply for this position. - Janet -------------------------------------------------------------------------- Senior Lecturer/Associate Professor in Neuroscience SCHOOL OF PSYCHOLOGY University of Queensland, Brisbane, Australia The School of Psychology is seeking applications for a continuing Senior Lecturer (Level C) or Associate Professor (Level D) position in any area of neuroscience, although preference may be given to an applicant in the field of cognitive neuroscience, and consideration will be given to the extent to which the applicant's research profile complements the research strengths of the School. The School of Psychology is one of the largest and most prestigious schools of psychology in Australia. It is internationally recognised for research strengths across the breadth of psychology - cognitive psychology, psychophysiology, and clinical neuropsychology are areas of particular research strength. Selection will be based primarily on research standing, teaching excellence, and capacity to attract and supervise postgraduate research students. Applicants should possess a PhD in psychology. At Level C, the successful appointee will have a developing international reputation for his or her research in neuroscience, have a strong research track record in relation to both publication and external research income, a reputation for high quality teaching, and be an experienced research supervisor in psychology. At Level D, a strong international research reputation, an outstanding research track record, a reputation for high quality teaching, and evidence of successful supervision of postgraduate students in psychology are required. The successful appointee will be expected to pursue a strong and productive program of research in neuroscience, to supervise psychology honours and postgraduate research theses in neuroscience, to contribute to the teaching of neuroscience in the undergraduate and honours teaching programs in the School, and to strengthen the links between the School and other centres and schools involved in neuroscience research at the University of Queensland. The remuneration package will be in the range $65,665- $75,716 per annum for Level C plus 17% employer superannuation contributions, and between $79,066- $87,106 per annum for Level D plus 17% employer superannuation contributions. Obtain the position description and selection criteria online or contact Ms Natasha Centis on +61 7 3365 6444 or n.centis at psy.uq.edu.au. Contact Professor Deborah Terry at d.terry at psy.uq.edu.au or telephone +61 7 3365 6220 to discuss the position in more detail. Send applications to the Principal Personnel Officer, Faculty of Social and Behavioural Sciences, University of Queensland, Brisbane, QLD 4072, Australia, or email j.laing at admin.uq.edu.au Closing date for applications: 23 September 2002 Reference Number: 3006530 -- Professor Deborah J. Terry Head of School School of Psychology University of Queensland Brisbane QLD 4072 Australia Phone: +61 7 3365 6220 Fax: +61 7 3365 4466 Email: deborah at psy.uq.edu.au From nnsp02 at neuro.kuleuven.ac.be Mon Jul 1 06:52:52 2002 From: nnsp02 at neuro.kuleuven.ac.be (Neural Networks for Signal Processing 2002) Date: Mon, 01 Jul 2002 12:52:52 +0200 Subject: 2002 IEEE International Workshop on Neural Networks for Signal Processing; Early Registration Deadline: July 31, 2002 Message-ID: <3D203484.734B21B3@neuro.kuleuven.ac.be> ------------------------------------------------------------------------- 2002 IEEE International Workshop on Neural Networks for Signal Processing September 4-6, 2002 Martigny, Valais, Switzerland Early registration deadline: July 31, 2002 ------------------------------------------- http://isp.imm.dtu.dk/nnsp2002 The twelfth in a series of IEEE NNSP workshops will be held in Martigny, Switzerland at the H=F4tel du Parc - Centre du Congres. The workshop will feature a strong technical program complemented by a series of exciting plenary talks, a special session on bioinformatics and a tutorial on "Nonlinear Adaptive Filtering for MIMO Multi-Antenna Wireless Communications" given by Prof. S.Y. Kung. All will be included in the registration package as well as special events such as a banquet in an old monastery and reception at the Gianadda museum. Please refer to the workshop web site for more information and note the early registration deadline of July 31, 2002. Plenary Talks: Faking unlabeled data for geometric regularization by Yoshua Bengio, University of Montreal, Canada Occam's Razor and Infinite Models by Zoubin Ghahramani, University College London, UK Hidden Markov models of proteins and DNA by Anders Krogh, Copenhagen University, Denmark From Peter.Flach at bristol.ac.uk Tue Jul 2 14:09:06 2002 From: Peter.Flach at bristol.ac.uk (Peter A. Flach) Date: Tue, 2 Jul 2002 19:09:06 +0100 Subject: MSc in Machine Learning and Data Mining at Bristol, UK Message-ID: *** please post and distribute as appropriate *** The Computer Science Department at the University of Bristol has been running an Advanced Computing MSc in Machine Learning and Data Mining since 1998. We are still accepting applications for enrolment in October 2002. See the URLs below for more information on the course and on how to apply. MSc Admissions: http://www.cs.bris.ac.uk/Admissions/msc.html MLDM MSc course: http://www.cs.bris.ac.uk/Teaching/MachineLearning/ Machine Learning research: http://www.cs.bris.ac.uk/Research/MachineLearning/ --Peter -- +------------------------------------+------------------------------------+ Peter A. Flach, Reader in Machine Learning Peter.Flach at bristol.ac.uk Dept. of Computer Science, University of Bristol tel. +44 117 954 5162 Merchant Venturers Building, Woodland Road fax +44 117 954 5208 Bristol BS8 1UB, United Kingdom http://www.cs.bris.ac.uk/~flach/ +------------------------------------+------------------------------------+ From dgw at MIT.EDU Fri Jul 5 11:46:21 2002 From: dgw at MIT.EDU (David Weininger) Date: Fri, 05 Jul 2002 11:46:21 -0400 Subject: book announcement--Rao Message-ID: <2002070511462130675@outgoing.mit.edu> I thought readers of the Connectionists List might be interested in this book. For more information, please visit http://mitpress.mit.edu/0262182246/ Thank you! Best, David Probabilistic Models of the Brain Perception and Neural Function edited by Rajesh P. N. Rao, Bruno A. Olshausen, and Michael S. Lewicki Neurophysiological, neuroanatomical, and brain imaging studies have helped to shed light on how the brain transforms raw sensory information into a form that is useful for goal-directed behavior. A fundamental question that is seldom addressed by these studies, however, is why the brain uses the types of representations it does and what evolutionary advantage, if any, these representations confer. It is difficult to address such questions directly via animal experiments. A promising alternative is to use probabilistic principles such as maximum likelihood and Bayesian inference to derive models of brain function. This book surveys some of the current probabilistic approaches to modeling and understanding brain function. Although most of the examples focus on vision, many of the models and techniques are applicable to other modalities as well. The book presents top-down computational models as well as bottom-up neurally motivated models of brain function. The topics covered include Bayesian and information-theoretic models of perception, probabilistic theories of neural coding and spike timing, computational models of lateral and cortico-cortical feedback connections, and the development of receptive field properties from natural signals. Rajesh P. N. Rao is Assistant Professor in the Department of Computer Science and Engineering at the University of Washington. Bruno A. Olshausen is Associate Professor in the Department of Psychology and the Center for Neuroscience at the University of California, Davis. Michael S. Lewicki is Assistant Professor in the Department of Computer Science and the Center for the Neural Basis of Cognition at Carnegie Mellon University. 8 x 10, 345 pp., cloth, ISBN 0-262-18224-6 Neural Information Processing series A Bradford Book ______________________ David Weininger Associate Publicist The MIT Press 5 Cambridge Center, 4th Floor Cambridge, MA 02142 617 253 2079 617 253 1709 fax http://mitpress.mit.edu From jls at cs.man.ac.uk Fri Jul 5 05:05:41 2002 From: jls at cs.man.ac.uk (Jon Shapiro) Date: Fri, 5 Jul 2002 10:05:41 +0100 Subject: Lectureship in Robotics; Manchester University Message-ID: <20020705100541.M1699@mortlach.cs.man.ac.uk> University of Manchester Department of Computer Science Lecturer in Robotics The Computer Science Department at the University of Manchester is seeking to appoint a lecturer in Robotics. The Department has recently been given a 5* rating in the Research Assessment Exercise. The Department has built up one of the UK leading mobile robotics labs, supported by a well-trained technical support staff. We currently teach both an undergraduate and an MSc course in mobile robotics which the successful candidate will be expected to take over. It is expected that the appointee will join a strong Artificial Intelligence group and contribute to or complement the current activities in machine learning, spatial and temporal reasoning, and computation of language. There will also be the opportunity to collaborate with other research groups within the Department. Of particular interest are the following: - the Imaging Science and Biomedical Engineering group, who are interested in medical robotics and computer-aided surgery; - the Mobile Systems Architecture group, who need a platform for research into terrain mapping and geolocation, using radio communication signals in a mobile Wireless LAN; - the Advanced Interfaces group, who are interested in using video footage captured by a robot camera to reverse engineer buildings. The successful candidate will have a track record of publications in these or related areas. More information about current robotics activity in the Department can be found at http://www.cs.man.ac.uk/robotics/ Annex A Particulars of Appointment - Lecturer in Robotics, Department of Computer Science Job Description The Lecturer's primary duties will lie in teaching, research, and administration. During the probationary period, the Lecturer will expect to be granted some relief from standard teaching and administrative duties in order to develop their research capabilities. Research The appointee will join a strong Artificial Intelligence group and will be expected to contribute to or complement the current research activities in machine learning, spatial and temporal reasoning, and computation of language. The appointee will undertake research in their specialised area of robotics in one of the UK leading mobile robotics labs, built up by the Department over a number of years and will be supported by a well-trained technical support staff. There will also be the opportunity to collaborate with other research groups within the Department. Of particular interest are the following: - the Imaging Science and Biomedical Engineering group, who are interested in medical robotics and computer-aided surgery; - the Mobile Systems Architecture group, who need a platform for research into terrain mapping and geolocation, using radio communication signals in a mobile Wireless LAN; - the Advanced Interfaces group, who are interested in using video footage captured by a robot camera to reverse engineer buildings. The successful candidate will be expected to produce high quality publications in these or related areas, attending relevant conferences and workshops and disseminating recent research results whenever possible. The appointee will also be expected to make applications for research grants and funding from relevant bodies. More information about current robotics activity in the Department can be found at http://www.cs.man.ac.uk/robotics/ Teaching The successful candidate will be expected to undertake teaching both at undergraduate and postgraduate level and to contribute to the development of Computer Science teaching. The appointee will take over the teaching of the current undergraduate and MSc modules in mobile robotics. The appointee will be expected to develop teaching material and learning experiences for students in the light of current educational practice, in particular by attending the University's training course for newly appointed lecturers and to participate in the planning and development of courses within the framework of Departmental and Faculty committees. Other teaching duties may include supervising laboratory and examples classes, conducting tutorials and supervise undergraduate and MSc projects. The appointee may also be required to undertake the supervision of research students. Administration The appointee will be expected to undertake Departmental administrative duties as agreed with the Head of Department. Person Specification Applicants for the lectureship will be expected to have: - a good honours degree in Computer Science (and preferably a PhD) - several years experience in a research environment, either in industry or academia - a strong publication record in the relevant area (see Job Description above) or, if coming from industry, an equivalent indication of output - the ability and willingness to teach in their specialised areas and to contribute to core computer science teaching - the ability and willingness to undertake normal administrative duties - ability to work as part of a team - good communication skills - preferably some research experience as part of a multidisciplinary team ----------------------------------------------------------------------- Jonathan Shapiro Computer Science Dept Room: 2.34 University of Manchester Phone: 44-(0)161 275 6253 Oxford Road Fax: 44-(0)161 275 6204 Manchester M13 9PL E-mail: jls at cs.man.ac.uk United Kingdom. FTP site: ftp.cs.man.ac.uk in directory /pub/ai/jls WWW site: http://www.cs.man.ac.uk/ai/jls/jls.html From ingber at ingber.com Fri Jul 5 12:33:56 2002 From: ingber at ingber.com (Lester Ingber) Date: Fri, 5 Jul 2002 12:33:56 -0400 Subject: Financial Engineer Position Message-ID: <20020705163356.GA18280@alumnus> If you have very strong credentials for the position described below, please email your resume to: Lester Ingber Director R&D Financial Engineer A disciplined, quantitative, analytic individual proficient in prototyping and coding (such as C/C++, Maple/Mathematica, or Visual Basic, etc.) is sought for financial engineering/risk:reward optimization research position with established Florida hedge fund (over two decades in the business and $1 billion in assets under management). A PhD in a mathematical science, such as physics, statistics, math, or computer-science, is preferred. Hands-on experience in the financial industry is required. Emphasis is on applying state-of-the-art methods to financial time-series of various frequencies. Ability to work with a team to transform ideas/models into robust, intelligible code is key. Salary: commensurate with experience, with bonuses tied to the individual's and the firm's performance. Selection Process All applicants will be reviewed, and a long list will be generated for phone interviews. Other applicants will not be contacted further. From these phone interviews, a short list will be generated for face-to-face interviews. Our estimated date for final selection will be September to October 2002. Start date for this position may range anywhere from immediately to six months thereafter, depending on both the candidate's and the firm's needs. All available information is posted on http://www.ingber.com/open_positions.html -- Prof. Lester Ingber ingber at ingber.com ingber at alumni.caltech.edu www.ingber.com www.alumni.caltech.edu/~ingber From pfbaldi at ics.uci.edu Sat Jul 6 11:59:54 2002 From: pfbaldi at ics.uci.edu (Pierre Baldi) Date: Sat, 6 Jul 2002 08:59:54 -0700 Subject: Postdoctoral positions in Bioinformatics and Machine Learning, Univeristy of California, Irvine Message-ID: <000001c22506$2b13bae0$be04c380@time-slice.ics.uci.edu> Several NIH-supported postdoctoral positions in the areas of Bioinformatics and Machine Learning are available in the Department of Information and Computer Science (www.ics.uci.edu) and the Institute for Genomics and Bioinformatics (www.igb.uci.edu) at the University of California, Irvine. Areas of particular interest: protein structure/function prediction, analysis of high-throughput array data (e.g. DNA microarray data), gene regulation, systems biology, and all areas of machine learning. Prospective candidates should apply with a cover letter, CV, and names and email addresses of 2-3 referees to be sent, preferably by email, to: pfbaldi at ics.uci.edu. The positions are available immediately and the duration of the appointments are typically 2 years with possibility of renewal. Relevant faculty in the department include: P. Baldi, D. Kibler, R. Lathrop, E. Mjolsness, and P. Smyth. From: esann To: "Connectionists at cs.cmu.edu" References: From bogus@does.not.exist.com Mon Jul 8 11:42:56 2002 From: bogus@does.not.exist.com () Date: Mon, 8 Jul 2002 17:42:56 +0200 Subject: Call for special sessions: ESANN'2003 European Symposium on Artificial Neural Networks Message-ID: ---------------------------------------------------- | | | ESANN'2003 | | | | 11th European Symposium | | on Artificial Neural Networks | | | | Bruges (Belgium) - April 23-24-25, 2003 | | | | First announcement & call for special sessions | ---------------------------------------------------- The ESANN'2003 conference European Symposium on Artificial Neural Networks) will take place in Bruges (Belgium), on April 23-25, 2003. Please have a look to http://www.dice.ucl.ac.be/esann for details. Proposals for special sessions are solicited. Details on special sessions (what are the role, advantages and duties of special session organizers) are available at http://www.dice.ucl.ac.be/esann/callforspecialsessions.htm. We try as much as possible to avoid multiple sendings of this call for papers; however please apologize if you receive this e-mail twice, despite our precautions. You will find below a short version of this call for special sessions. What are special sessions ? --------------------------- A "special session" is a session that focuses on a particular topic in the artificial neural networks field. It is organized by a renowned scientist who solicits (but does not "invite") potential contributors. During the conference, the session begins with an introduction (or tutorial) to the session topic by the session organizer, making it easier for the participants to follow the subsequent talks. The following are examples of special sessions organized during ESANN'98 to 2002 conferences: Radial basis networks Neural networks for control Self-organizing maps for data analysis ANN for speech processing Cellular neural networks ANN for the processing of facial information Adaptive computation of data structures Remote sensing spectral image analysis Support vector machines Information extraction using unsupervised neural networks Spiking neurons Artificial neural networks and robotics Neural networks in medicine Time series prediction ANN for energy management systems Dedicated hardware implementations: perspectives on systems and applications Novel neural transfer functions ANN and evolutionary/genetic algorithms: hybrid approaches ANN in finance ANN and early vision processing Perspectives on Learning with Recurrent Networks Representation of high-dimensional data Neural Network Techniques in Fault Detection and Isolation Hardware and Parallel Computer Implementations of Neural Networks Exploratory Data Analysis in Medicine and Bioinformatics Neural Networks and Cognitive Science Who can organize special sessions ? ----------------------------------- Any scientist in the neural networks field is welcome to submit his/her proposal to organize a special session. He/she should first check is he/she is able to contact enough potential contributors in order to organize a sound session. How to organize special sessions ? ---------------------------------- We see the role of a session organizer as follows: to contact about 7-10 potential authors among specialists in the field and to ask them to submit a paper to the conference; to make sure that each paper is evaluated by 3 reviewers (including the session organizer); according to the reviews, to decide the final contents of the session, in agreement with the program committee; to chair the session during the conference; to present a tutorial (typically 30 minutes) at the beginning of the session. This tutorial should both introduce the topic to non-specialists, and make the link between the state-of-the-art and the papers presented in the session. to write a paper for the proceedings, with the contents of the tutorial. The following aspects must be taken into account: The session can include oral and poster presentations. Posters presentations are introduced by a 2-minutes spotlight during the oral session. There should be between 4 and 6 oral presentations during the session (including the introduction/tutorial), and between 0 and 6 posters. Soliciting contributors does NOT mean that their contributions are automatically accepted. Solicited contributions are submitted to a review process and will be rated according to their scientific value. This must be clearly explained by the session organizer when he/she first contacts the potential contributors. Soliciting contributors does NOT mean too that these contributors are invited to the conference. Even in case of acceptation of their submission, they will have to register to the conference and pay the registration fee, as any other participant. This must be clearly explained too by the session organizer when he/she first contacts the potential contributors. What are the advantages for special session organizers ? -------------------------------------------------------- Session organizers benefit from a 50% reduction of their own registration fees. However, travel and subsistence expenses must be taken in charge by the session organizers. Their article in the ESANN proceedings are limited to 12 pages instead of 6 pages. Depending on public funding obtained to organize the conference, the hotel costs of session organizers could be taken in charge. Nevertheless, we will not have the confirmation of this possibility before having to decide which special sessions will be organized. Those who submit a proposal to organize a special session must be aware that they could have to take in charge their hotel expenses. Deadlines --------- August 1, 2002: submission of proposals to organize a special session August 15, 2002: selection of special sessions for ESANN'2002 September 1, 2002: availability of the ESANN'2002 call for papers on this Web site end of September, 2002: contacts (by session organizer) with potential contributors December 6, 2002: deadline submission of papers (including special sessions) around December 15, 2002: dispatching of submissions to reviewers January 17, 2003: end of review process January 29, 2003: decision on the contents of the special sessions February 6, 2003: notification of acceptance/reject of submissions February 28, 2003: deadline for final versions of papers (including the introduction/tutorial by the session organizer) How to submit a proposal to organize a special session ? -------------------------------------------------------- Simply send an e-mail to the conference secretariat with the title of the session a brief description (+- 10 lines, to be published on the ESANN Web site) your affiliation, address, phone and fax number, e-mail address a description of your expertise in the field of the session (or a link to your personal Web page where this information can be found) Proposals are expected until August 1, 2002. Special sessions to be organized during the ESANN'2003 conference will be selected around August 15, 2002. Please mention in your e-mail if we will be able to contact you between August 1 and August 15, 2002. ======================================================== ESANN - European Symposium on Artificial Neural Networks http://www.dice.ucl.ac.be/esann * For submissions of papers, reviews,... Michel Verleysen Univ. Cath. de Louvain - Microelectronics Laboratory 3, pl. du Levant - B-1348 Louvain-la-Neuve - Belgium tel: +32 10 47 25 51 - fax: + 32 10 47 25 98 mailto:esann at dice.ucl.ac.be * Conference secretariat d-side conference services 24 av. L. Mommaerts - B-1140 Evere - Belgium tel: + 32 2 730 06 11 - fax: + 32 2 730 06 00 mailto:esann at dice.ucl.ac.be ======================================================== From M.Herbster at cs.ucl.ac.uk Tue Jul 2 17:14:33 2002 From: M.Herbster at cs.ucl.ac.uk (Mark HERBSTER) Date: Tue, 02 Jul 2002 22:14:33 +0100 Subject: Lectureship in Intelligent Systems at University College London Message-ID: <21036.1025644473@cs.ucl.ac.uk> UNIVERSITY COLLEGE LONDON Department of Computer Science Lectureship in Intelligent Systems We are seeking talented researchers who are keen to undertake innovative research embracing the application and theory of Intelligent Systems. The successful candidate is likely to already be working in Intelligent Systems or Machine Learning either in a leading university research centre or company. He/she will be expected to play a leading role in the research and teaching of Machine Learning in the Department, and especially in the development of the new MSc programme in Intelligent Systems. Applicants will have a higher degree in a relevant discipline and a strong Computer Science background. Experience of applying Machine Learning to business or industrial applications would be an advantage. The appointment will be on the Lecturer A or Lecturer B scale (22,604 to 34,671 pa including London Allowance) or on the Senior Lecturer scale (36,292 to 40,737 including London Allowance) according to experience. Further information regarding this post and the application form can be found on our website (www.cs.ucl.ac.uk/vacancies). Informal enquiries maybe directed to Professor Steve Wilbur (s.wilbur at cs.ucl.ac.uk). The closing date for applications is Tuesday, 30th July 2002. --- Mark Herbster Department of Computer Science University College London Gower Street London WC1E 6BT United Kingdom Phone: +44 (0)20 7679 3684 Fax: +44 (0)20 7387 1397 From qian at brahms.cpmc.columbia.edu Tue Jul 9 12:16:21 2002 From: qian at brahms.cpmc.columbia.edu (Ning Qian) Date: Tue, 9 Jul 2002 12:16:21 -0400 Subject: paper: structure from motion Message-ID: <200207091616.g69GGLA09177@brahms.cpmc.columbia.edu> Dear colleagues, The following paper (1.3MB) is available at: http://brahms.cpmc.columbia.edu/publications/sfm.pdf (It was published a couple of months ago but I never announced it before.) Computing relief structure from motion with a distributed velocity and disparity representation Julian Martin Fernandez, Brendon Watson, and Ning Qian, Vision Research, 2002, 42:883-898. Recent psychophysical experiments suggest that humans can only recover relief structure from motion (SFM); i.e., an object's 3D shape can only be determined up to a stretching transformation along the line of sight. Here we propose a physiologically plausible model for the computation of relief SFM, which is also applicable to the related problem of motion parallax. We assume that the perception of depth from motion is related to the firing of a subset of MT neurons tuned to both velocity and disparity. The model MT neurons are connected to each other laterally to form modulatory interactions. The overall connectivity is such that when a zero-disparity velocity pattern is fed into the system, the most responsive neurons are not those tuned to zero disparity, but instead are those having preferred disparities consistent with the relief structure of the velocity pattern. The model computes the correct relief structure under a wide range of parameters and can also reproduce the SFM illusions involving coaxial cylinders. It is consistent with the psychophysical observation that subjects with stereo impairment are also deficient in perceiving motion parallax, and with the physiological data that the responses of direction- and disparity-tuned MT cells covary with the perceived surface order of bistable SFM stimuli. Many earlier papers at http://brahms.cpmc.columbia.edu/ are now also available in PDF format. Best regards, Ning ------------------------------------------------------------- http://brahms.cpmc.columbia.edu Ning Qian, Ph. D. qian at brahms.cpmc.columbia.edu Associate Professor nq6 at columbia.edu Ctr. Neurobiology & Behavior Columbia University / NYSPI 212-543-5213 (Office) Kolb Annex Rm 730 212-543-5161 (Lab/fax) 722 W. 168th Street 212-543-6048 (Lab) New York, NY 10032, USA ------------------------------------------------------------- From jose at psychology.rutgers.edu Tue Jul 9 01:53:09 2002 From: jose at psychology.rutgers.edu (stephen j hanson) Date: 09 Jul 2002 01:53:09 -0400 Subject: Systems Admin/ Research Staff Position --Cognitive Science/Cognitive Neuroscience/Computation Message-ID: <1026194002.1980.47.camel@vaio> Systems Adminstration/UNIX/LINUX--Research Staff IMMEDIATE OPENING-- 08/01/02 PSYCHOLOGY DEPARTMENT-RUTGERS UNIVERSITY--Newark Campus-- We are searching for an individual who can adminster the computing resources of the Psychology Department at Rutgers (Newark) Resources include a network of Sun/LINUX workstations, PCs and Macs, printers, pc-voice mail system and various devices (scanners, projecters etc..). The individual will be responsible for installing and debugging software, and various routine system administration activites. About half their time will be spent in research involving Cognitive Neuroscience/Cognitive Science especially related to Connectionist networks (or Neural Networks and Computational Neuroscience. Familiarity with C programming, UNIX system internals (BSD, System V, Solaris, Linux) and Windows (XX, NT) and local area networks running TCP/IP is required. Image processing or graphics programing experience are pluses. Candidates should possess either a BS/MS in Computer Science, Cognitive Science, AI or other relevant fields or equivalent experience. We are located 15-20 minutes from Downtown Manhattan and are within minutes of the NJPAC in the heart of Newark. Salary is competitive and will be dependent upon qualifications and experience. Rutgers University is an equal opportunity affirmative action employer. Please send resumes and references to Stephen J. Hanson Department of Psychology RUMBA Labs 101 Warren Street Rutgers University Newark, New Jersey, 07102 Direct email inquiries or resumes to: jose at psychology.rutgers.edu Please indicate on SUBJECT Line: SYS ADM as a keyword. From cowen at mail.nsma.arizona.edu Thu Jul 11 20:34:50 2002 From: cowen at mail.nsma.arizona.edu (cowen@mail.nsma.arizona.edu) Date: Thu, 11 Jul 2002 17:34:50 -0700 (MST) Subject: Job Posting, Message-ID: <1026434090.3d2e242ab0db4@email.arl.arizona.edu> Job No. 24458 Appointed Position Job Title: Associate Research Scientist Department: Neural Systems Memory and Aging Division of Arizona Research Laboratories Salary: DOE Benefits Hours: 40 per week Opening: 7/10/02 Closing: 7/17/02 Position Summary: The Arizona Research Laboratories Division of Neural Systems, Memory and Aging seek candidates for the position of Associate Research Scientist. The successful candidate will be part of a collaborative research project between the University of Arizona and University of California at Davis. The primary worksite will be at the California Regional Primate Research Center in Davis, CA. This is an extended temporary position of approximately six months or more duration. This individual will be responsible for primate electrophysiology experiments using an advanced parallel recording systems developed here in Tucson (the Neuralynx Cheetah system). Minimum Qualifications: Ph.D. in Neuroscience, Behavioral Psychology or a related field. Preferred Qualifications: * Experience in neuroscience, with an emphasis in neural systems. * Familiarity with the Neuralynx Cheetah Data Acquisition System. * Experience in computer data acquisition and analysis, including Matlab. To apply, please submit a cover letter, resume and the names and contact information for three references to: Search Committee c/o Luann Snyder ARL Division of Neural Systems, Memory and Aging P.O. Box 245115 The University of Arizona Tucson, AZ 85724-5115 Please reference job number 24458. For consideration, complete requested documentation must be received by midnight of the closing date. The University of Arizona is an EEO/AA Employer-M/W/D/V From m.niranjan at dcs.shef.ac.uk Fri Jul 12 09:24:01 2002 From: m.niranjan at dcs.shef.ac.uk (Mahesan Niranjan) Date: Fri, 12 Jul 2002 14:24:01 +0100 (BST) Subject: Job Job Job Message-ID: The University of Sheffield has three Faculty openings (Lecturer / Senior Lecturer) in Computer Science. One of these is in "any area of CS" and two in "bioinformatics". Both definitions are sufficiently broad to be of interest to subscribers of this forum. In the Bioinformatics area, we have excellent collaborations with local developmental biologists and medics. http://www.shef.ac.uk/jobs/adcjobs Academic positions in the UK offer excellent freedom to pursue ideas that interest you. Other features of the job include low pay, poor student motivation, vague concept of tenure, periodic assessment of teaching and research quality in departments... If you know someone who might be interested, please point them in my direction. thanx niranjan ____________________________________________________________________ Mahesan Niranjan Phone: 44 114 222 1805 Professor of Computer Science FaX: 44 114 222 1810 University of Sheffield Email: M.Niranjan at dcs.shef.ac.uk http://www.dcs.shef.ac.uk/~niranjan ____________________________________________________________________ From marcus.hutter at gmx.net Sat Jul 13 05:39:18 2002 From: marcus.hutter at gmx.net (Marcus Hutter) Date: Sat, 13 Jul 2002 11:39:18 +0200 Subject: PhD & Postdoc Position Available Message-ID: <010901c22a51$8036fc80$78f536cb@ho> PhD & Postdoc Position Available ----------------------------------------- IDSIA, Switzerland, is seeking for one outstanding PhD student and one PostDoc with excellent mathematical skills interested in reinforcement learning, algorithmic information theory, Kolmogorov complexity, Minimal Description Length, computational complexity theory, information theory and statistics, universal Solomonoff induction, universal Levin search, sequential decision theory, adaptive control theory, and/or related areas. Possible backgrounds are computer science, physics, mathematics, etc. The initial appointment will be for 2 years. Normally there will be a prolongation. The new PhD student/PostDoc will interact with Marcus Hutter and Juergen Schmidhuber and other people at IDSIA. See http://www.idsia.ch/~marcus/idsia/phdpos1.htm for more information on the PhD position and http://www.idsia.ch/~marcus/idsia/postdoc1.htm for more information on the PostDoc position. Applicants should submit: (i) Detailed curriculum vitae, (ii) List of three references and their email addresses, (iii) Concise statement of their research interests (two pages max). Please send all documents to: Marcus Hutter, IDSIA, Galleria 2, 6928 Manno (Lugano), Switzerland. Applications can also be submitted by email to marcus at idsia.ch (2MB max). WWW pointers to ps/pdf/doc/html files are welcome. Use Firstname.Lastname.DocDescription.DocType for filename convention. Thanks for your interest Marcus Hutter, senior researcher, IDSIA Istituto Dalle Molle di Studi sull'Intelligenza Artificiale Galleria 2 CH-6928 Manno(Lugano) - Switzerland Phone: +41-91-6108668 Fax: +41-91-6108661 E-mail marcus at idsia.ch http://www.idsia.ch/~marcus --------------------------------------------------------------------------- ABOUT IDSIA. Our research focuses on artificial neural nets, reinforcement learning, complexity and generalization issues, unsupervised learning and information theory, forecasting, artificial ants, combinatorial optimization, evolutionary computation. IDSIA is small but visible, competitive, and influential. IDSIA's algorithms hold the world records for several important operations research benchmarks (see Nature 406(6791):39-42 for an overview of artificial ant algorithms developed at IDSIA). In the "X-Lab Survey" by Business Week magazine, IDSIA was ranked in fourth place in the category "COMPUTER SCIENCE - BIOLOGICALLY INSPIRED" - after the Santa Fe Institute, Stanford University, and EPFL (also in Switzerland). Its comparatively tiny size notwithstanding, IDSIA also ranked among the top ten labs worldwide in the broader category "ARTIFICIAL INTELLIGENCE". IDSIA is located near the beautiful city of Lugano in Ticino (pictures), the scenic southernmost province of Switzerland, origin of special relativity and the WWW. Milano, Italy's center of fashion and finance, is 1 hour away, Venice 3 hours. Our collaborators at CSCS (the Swiss supercomputing center) are right beneath us; we are also affiliated with the University of Lugano and SUPSI. Switzerland boasts the highest citation impact factor, the highest supercomputing capacity pc (per capita), the most Nobel prizes pc (450% of the US value), and perhaps the best chocolate. From jose at psychology.rutgers.edu Tue Jul 16 08:32:24 2002 From: jose at psychology.rutgers.edu (Stephen J. Hanson) Date: 16 Jul 2002 08:32:24 -0400 Subject: POSTDOC POSITION--IMMEDIATE OPENING Message-ID: <1026822750.3327.140.camel@madison> COGNITIVE/COMPUTATIONAL NEUROSCIENCE POSTDOCTORAL POSITION at RUTGERS UNIVERSITY, Newark Campus. The Rutgers University Mind/Brain Analysis (RUMBA) Project anticipates making one postdoctoral appointment, which is to begin in the FALL (August/September) of 2002. This positions are for a minimum of 2 years, with the possibility of continuation for 1 more year and will be in the areas of specialization of cognitive neuroscience with emphasis on the development of new paradigms and methods in neuroimaging, mathematical modeling, signal processing or data analysis in functional brain imaging. Particular interest is in methods and algorithms for fusion of EEG/fMRI. Applications are welcomed begining immediately and review will continue until the position is filled. Rutgers University is an equal opportunity/affirmative action employer. Qualified women and minority candidates are especially encouraged to apply. Send CV and three letters of recommendation and 1 reprint to Professor S.J. Hanson, Department of Psychology, Rutgers University, Newark, NJ 07102. Email enquiry can be made to jose at psychology.rutgers.edu please put "RUMBA POSTDOC" in your subject field also see http://www.rumba.rutgers.edu. From Zoubin at gatsby.ucl.ac.uk Wed Jul 17 07:44:58 2002 From: Zoubin at gatsby.ucl.ac.uk (Zoubin Ghahramani) Date: Wed, 17 Jul 2002 12:44:58 +0100 (BST) Subject: Faculty Position in Intelligent Systems / Machine Learning in London Message-ID: <200207171144.MAA07760@cajal.gatsby.ucl.ac.uk> For those of you who may not have seen this ad for a faculty position in Intelligent Systems in the Department of Computer Science at University College London (UCL), I've enclosed it below. I'd like to add that the MSc in Intelligent Systems was established and is taught jointly by the Computer Science department and the Gatsby Unit at UCL (see http://www.gatsby.ucl.ac.uk). We hope that the successful applicant for this post will have strong interactions with members of the Gatsby Unit, adding to and complementing our key strengths in learning theory, Bayesian methods, reinforcement learning, pattern recognition, kernel machines and graphical models. -Zoubin Ghahramani University College London ---------------------------------------------------------------------- UNIVERSITY COLLEGE LONDON Department of Computer Science We are seeking talented researchers who are keen to undertake innovative research embracing the application and theory of Intelligent Systems. The successful candidate is likely to already be working in Intelligent Systems either in a leading university research centre or company. He/she will be expected to play a leading role in the research and teaching of Intelligent Systems in the Department, and especially in the development of the new MSc programme in Intelligent Systems. Applicants will have a higher degree in a relevant discipline and a strong Computer Science background. Experience of applying Intelligent Systems to business or industrial applications would be an advantage. The appointment will be on the Lecturer A or Lecturer B scale (?22,604 to ?34,671 pa including London Allowance) according to experience. Further information regarding this post and the application form can be found on our website (http://www.cs.ucl.ac.uk/vacancies). Informal enquiries may be directed to Professor Steve Wilbur (s.wilbur at cs.ucl.acuk). The closing date for applications is Tuesday, 30th July 2002. Taking Action for Equality Claire Cawley Human Resources Administrator and PA to Professor Wilbur, Head of Department Department of Computer Science University College London Gower Street London WC1E 6BT T: 020 7679 3676 (internal: 33676) F: 020 7387 1397 E: c.cawley at cs.ucl.ac.uk I: www.cs.ucl.ac.uk/staff/C.Cawley/ From niki at cis.ohio-state.edu Wed Jul 17 14:38:51 2002 From: niki at cis.ohio-state.edu (Nicole Roman) Date: Wed, 17 Jul 2002 14:38:51 -0400 Subject: Tech report on location-based segregation Message-ID: <3D35B9BB.23CB64BD@cis.ohio-state.edu> Dear Colleagues, It is my pleasure to announce the availability of the following technical report. Thanks for your attention, Nicoleta Roman ************************************ "Speech segregation based on sound localization", Technical Report #16, June 2002. Department of Computer and Information Science The Ohio State University Nicoleta Roman, The Ohio State University DeLiang Wang, The Ohio State University Guy J. Brown, University of Sheffield ************************************* Abstract --------- At a cocktail party, we can selectively attend to a single voice and filter out all the other acoustical interferences. How to simulate this perceptual ability remains a great challenge. This paper describes a novel machine learning approach to speech segregation, in which a target speech signal is separated from interfering sounds using spatial location cues: interaural time differences (ITD) and interaural intensity differences (IID). The auditory masking effect motivates the notion of an ?ideal? time-frequency binary mask, which selects the target if it is stronger than the interference in a local time-frequency (T-F) unit. We observe that within a narrow frequency band, modifications to the relative strength of the target source with respect to the interference trigger systematic deviations for ITD and IID. For a given spatial configuration, this interaction produces characteristic clustering in the binaural feature space. Consequently, we perform pattern classification in order to estimate ideal binary masks. A systematic evaluation shows that the resulting system produces masks very close to ideal binary ones, and gives a significant improvement in performance over an existing approach, as quantified by changes in signal-to-noise ratio before and after segregation. ************************************** The manuscript is available for download at: ftp://ftp.cis.ohio-state.edu/pub/tech-report/2002/TR16.pdf Related sound demos can be found at: http://www.cis.ohio-state.edu/~niki/soundemo.html A preliminary version of this work is included in the Proceedings of 2002 ICASSP. From arbib at pollux.usc.edu Wed Jul 17 21:53:27 2002 From: arbib at pollux.usc.edu (Michael Arbib) Date: Thu, 18 Jul 2002 09:53:27 +0800 Subject: Post-Doctoral Traineeship in Computational and Cognitive Neuroscience at the University of Southern California Message-ID: <200207180155.JAA06657@bilby.cs.uwa.edu.au> The Neuroscience Program at the University of Southern California (http://www.usc.edu/dept/nbio/ngp/) has an active program of research in a broad range of areas of Computational and Cognitive Neuroscience including vision, motor control, visuomotor coordination, linguistics (with an active new focus on an action-oriented approach to the evolution of brain mechanisms of language), neuroinformatics, neural engineering, and memory and learning. The Training Grant in Computational and Cognitive Neuroscience at USC has one opening for a postdoctoral trainee, effective immediately. Candidates must be US nationals or permanent residents. In addition to the NIH-mandated salary, the trainee will receive 4 units/year tuition, payment of mandatory student fees, a travel allowance of $500 and $3500 for training related expenses (M&S). Candidates with a strong interest in Computational and/or Cognitive Neuroscience should email an application to Michael Arbib (arbib at pollux.usc.edu) with the following materials: 1. Name, email, and statement of US citizenship or permanent resident status 2. Undergraduate institution, major, GPA and GRE. 3. Ph.D. thesis title and abstract, institution and years of study, list of all graduate courses taken (with grades). 4. A one page statement of interest in Computational and/or Cognitive Neuroscience as an area for postdoctoral research, with a list of one or more possible supervisors at USC with reasons for your choice(s). 5. A list of publications, presentations, and awards. In addition, the applicant should ask three researchers with relevant expertise to send Dr. Arbib a half-page email attesting to the applicant's suitability for a traineeship. Applications and support letters must be received by August 19, 2002 for consideration for Traineeships for 12 months beginning either September 1, 2002 or January 1, 2003 (candidates should indicate for which of these periods they are available). ******************************************************* Michael A. Arbib USC Brain Project University of Southern California Los Angeles, CA 90089-2520 Phone (213) 740-9220 FAX (213) 740-5687 arbib at pollux.usc.edu [Additional details for Express Mail: Hedco Neuroscience Building, Room 5, 3614 Watt Way; Phone Contact 213-740-1176.] ______________________ http://www-hbp.usc.edu/people/arbib.htm http://www-hbp.usc.edu/ http://mitpress.mit.edu/e-books/HBTNN2e http://www.cs.usc.edu/ ******************************************************* From d.mareschal at bbk.ac.uk Wed Jul 17 12:52:01 2002 From: d.mareschal at bbk.ac.uk (Denis Mareschal) Date: Wed, 17 Jul 2002 17:52:01 +0100 Subject: 3 YEAR PHD POSITION in Belgium Message-ID: Please forward this to any interested parties. Please DO NOT REPLY DIRECTLY TO ME. cheers, Denis ============================ 3 year PHD Position A 3-year position for a graduate student in cognitive science, funded by a research grant from the European Commission, is available immediately at the Free University of Brussels (ULB)in the laboratory of Axel Cleeremans. Dr. Cleeremans' lab is primarily in the area of implicit learning (i.e., learning without being consciously aware of learning having taken place), adn consciousness. The research carried out by his group is cutting edge and has been published in some of the foremost international journals in the world (e.g., Nature Neuroscience). The research in this lab uses a combination of neural network modelling and experimental testing to better understand observable behavioral phenomena. In the area of connectionist modeling of psychological processes, Cleeremans' lab is one of the best known in Europe. It is comprised of a relatively small and tight-knit group of researchers, and is well funded and well equipped. Details of the research activities in this lab can be found at : http://srsc.ulb.ac.be/axcWWW/axc.html The European Commission funding for a qualified graduate student will run until February 2005. The candidate will be expected to enroll in the Ph.D. program at ULB, under Dr. Cleermans' supervision. A background in experimental psychology and a good grounding in computer programming (e.g., C++, JAVA, Delphi, or similar languages) are the only pre-requisites for this position. Funding is available as of September 1st. Applications will continue to be taken an appropriate candidate is found. Some funding restrictions apply. The candidate must be less than 35 years old, must be a citizen (or long temr resident) of one of the member or associate-member states of the European Union. However, cannot be a Belgian citizen and cannot have resided in Belgium for more than 1 year in the last two years. For further information, please contact: - Dr. Axel Cleeremans directly: axcleer at ulb.ac.be, - Dr. Robert French, at the University of Liege in Belgium, director of the EC project: rfrench at ulg.ac.be For more information on the EU research project, and the partners in the project, please consult the following URL's: http://www.ulg.ac.be/cogsci/bmlf/ and for a detailed description of the Project itself: http://www.ulg.ac.be/rfrench/bmlf.pdf ================================================= Dr. Denis Mareschal Centre for Brain and Cognitive Development School of Psychology Birkbeck College University of London Malet St., London WC1E 7HX, UK tel +44 020 7631-6582/6207 fax +44 020 7631-6312 http://www.psyc.bbk.ac.uk/staff/dm.html ================================================= From juergen at idsia.ch Thu Jul 18 10:18:20 2002 From: juergen at idsia.ch (Juergen Schmidhuber) Date: Thu, 18 Jul 2002 16:18:20 +0200 Subject: near-optimal computable predictions Message-ID: <3D36CE2C.E3CD6023@idsia.ch> The Speed Prior: a new simplicity measure yielding near-optimal computable predictions (Juergen Schmidhuber, IDSIA) In J. Kivinen and R. H. Sloan, eds, Proc. 15th Annual Conf. on Computational Learning Theory (COLT), 216-228, Springer, 2002; based on section 6 of http://arXiv.org/abs/quant-ph/0011122 (2000) http://www.idsia.ch/~juergen/speedprior.html ftp://ftp.idsia.ch/pub/juergen/colt.ps Solomonoff's optimal but noncomputable method for inductive inference assumes that observation sequences x are drawn from an recursive prior distribution mu(x). Instead of using the unknown mu(x) he predicts using the celebrated universal enumerable prior M(x) which for all x exceeds any recursive mu(x), save for a constant factor independent of x. The simplicity measure M(x) naturally implements "Occam's razor" and is closely related to the Kolmogorov complexity of x. However, M assigns high probability to certain data x that are extremely hard to compute. This does not match our intuitive notion of simplicity. Here we suggest a more plausible measure derived from the fastest way of computing data. In absence of contrarian evidence, we assume that the physical world is generated by a computational process, and that any possibly infinite sequence of observations is therefore computable in the limit (this assumption is more radical and stronger than Solomonoff's). Then we replace M by the novel Speed Prior S, under which the cumulative a priori probability of all data whose computation through an optimal algorithm requires more than O(n) resources is 1/n. We show that the Speed Prior allows for deriving a computable strategy for optimal prediction of future y, given past x. Then we consider the case that the data actually stem from a nonoptimal, unknown computational process, and use Hutter's recent results to derive excellent expected loss bounds for S-based inductive inference. Assuming our own universe is sampled from S, we predict: it won't get many times older than it is now; large scale quantum computation won't work well; beta decay is not random but due to some fast pseudo-random generator which we should try to discover. Juergen Schmidhuber http://www.idsia.ch/~juergen From levy at cs.brandeis.edu Thu Jul 18 06:10:37 2002 From: levy at cs.brandeis.edu (Simon Levy) Date: Thu, 18 Jul 2002 06:10:37 -0400 Subject: Ph.D. thesis announcement: Infinite RAAM Message-ID: <3D36941D.2090707@cs.brandeis.edu> Levy, Simon D. (2002). Infinite RAAM: Initial Investigations into a Fractal Basis for Cognition. Ph.D. Thesis, Brandeis University, July 2002. Abstract This thesis attempts to provide an answer to the question ``What is the mathematical basis of cognitive representations?'' The answer we present is a novel connectionist framework called Infinite RAAM. We show how this framework satisfies the cognitive requirements of systematicity, compositionality, and scalable representational capacity, while also exhibiting ``natural'' properties like learnability, generalization, and inductive bias. The contributions of this work are twofold: First, Infinite RAAM shows how connectionist models can exhibit infinite competence for interesting cognitive domains like language. Second, our attractor-based learning algorithm provides a way of learning structured cognitive representations, with robust decoding and generalization. Both results come from allowing the dynamics of the network to devise emergent representations during learning. An appendix provides Matlab code for the experiments described in the thesis. Keywords: Neural Networks, Fractals, Connectionism, Language, Grammar. Postscript: http://www.demo.cs.brandeis.edu/papers/levythesis.ps Gzipped: http://www.demo.cs.brandeis.edu/papers/levythesis.ps.gz PDF: http://www.demo.cs.brandeis.edu/papers/levythesis.pdf From ckiw at dai.ed.ac.uk Fri Jul 19 09:24:54 2002 From: ckiw at dai.ed.ac.uk (Chris Williams) Date: Fri, 19 Jul 2002 14:24:54 +0100 (BST) Subject: Faculty Positions at the University of Edinburgh Message-ID: The School of Informatics invites applications from candidates of international standing for three appointments at the level of either Reader or Lecturer. Outstanding candidates in all the following areas are invited to apply: Algorithms; Cognitive Systems; Computer Vision, Graphics and Robotics; Computing Systems Architecture; Knowledge Representation and Reasoning; Machine Learning and Probabilistic Modelling; and Software Engineering. [The Lecturer and Reader posts are roughly equivalent to US Assistant and Associate Professor levels.] We invite applicants from all areas of machine learning and probabilistic modelling. We have particular interest in probabilistic graphical models, learning applied to computer vision, and data mining of complex data types. Research in machine learning and probabilistic modelling occurs in many areas of the School, with the Institute for Adaptive and Neural Computation acting as a hub for these activities. These research activities are supported by collaborations within and outwith the university, for example with the Department of Child Life and Health (condition monitoring on premature babies) and with the Royal Observatory of Edinburgh (astronomical data mining). Informal enquiries with regard to Machine Learning and Probabilistic Modelling may be made to Dr Chris Williams (c.k.i.williams at ed.ac.uk). Informatics at Edinburgh is one of the top-ranked departments in the United Kingdom. Candidates should demonstrate a world-class research record and both interest and ability in teaching. Candidates for a readership will be expected to demonstrate the ability to take on research leadership in their respective area. Informal enquiries to Bonnie Webber, +44 131 650 4190 (bonnie.webber at ed.ac.uk) or to Michael Fourman, +44 131 650 2703 (hod at informatics.ed.ac.uk). Further information can be found at both http://www.informatics.ed.ac.uk/events/vacancies/ and http://www.jobs.ed.ac.uk/. Salary scale: Lecturer 20,470 - 32,537 pounds p.a. (under review) Reader 34,158 - 38,603 pounds p.a. (under review) Please quote Ref: 311616 Letters of application should include a curriculum vitae and the names and addresses of 3 referees. Please include fax numbers and email addresses for referees if possible. Applications should be addressed to Division of Informatics (c/o Ms. Eleanor Kerse), University of Edinburgh, and sent, to arrive not later than 23 August 2002, by post (80 South Bridge, Edinburgh EH1 1HN, UK), fax (+44 (0)131 650 6516), or email (hod at informatics.ed.ac.uk). Applications can be made on-line through http://www.jobs.ed.ac.uk/. Closing date: 23 August 2002 Dr Chris Williams ckiw at dai.ed.ac.uk Institute for Adaptive and Neural Computation Division of Informatics, University of Edinburgh 5 Forrest Hill, Edinburgh EH1 2QL, Scotland, UK fax: +44 131 650 6899 tel: (direct) +44 131 651 1212 (department switchboard) +44 131 650 3090 http://www.dai.ed.ac.uk/homes/ckiw/ From mm at santafe.edu Mon Jul 22 15:18:43 2002 From: mm at santafe.edu (Melanie Mitchell) Date: Mon, 22 Jul 2002 13:18:43 -0600 (MDT) Subject: Graduate Research Positions Message-ID: <200207221918.g6MJIhv08743@taos.santafe.edu> I have openings for two graduate research assistants to work on a computer model of analogy-making between visual images. This work will build on the "Copycat" model of Hofstadter and Mitchell and will incorporate other approaches to high-level perception and image understanding, including those inspired by the field of "complex adaptive systems". More information about the project can be found at http://www.santafe.edu/~mm/analogy-vision.html. A recent paper describing the Copycat model, "Analogy-making as a complex adaptive system", can be downloaded from http://www.santafe.edu/~mm/paper-abstracts.html#amcas. Applicants must be willing to pursue a graduate degree in Computer Science and Engineering at the OGI School of Science and Engineering, Oregon Health & Science University, near Portland, Oregon, where I will be joining the faculty. The department web pages can be found at http://www.cse.ogi.edu. Proficiency in C, C++, or another high-level programming language is required. Background in cognitive science, psychology, computer science, mathematics, image processing and computer vision, and/or biology would be helpful. The assistanceship will cover tuition and stipend. To apply, send a resume with your research interests, list of relevant course work or experience, programming experience and languages, and any other information you think would be relevant, and the names and contact information of at least two professors or scientists who will act as references. Please send this information in electronic form to mm at santafe.edu. Applications will be considered until the positions are filled. Students of any nationality may apply. OGI is an equal opportunity employer and particularly welcomes applications from women and minority candidates. ----------------------------------- Melanie Mitchell Associate Professor Department of Computer Science and Engineering OGI School of Science & Engineering Oregon Health & Science University 20000 NW Walker Road Beaverton, OR 97006 E-mail: mm at santafe.edu From hadley at cs.sfu.ca Thu Jul 25 15:48:47 2002 From: hadley at cs.sfu.ca (Bob Hadley) Date: Thu, 25 Jul 2002 12:48:47 -0700 (PDT) Subject: Systematicity & Fallacies: Boden & Niklasson Message-ID: <200207251948.g6PJmll17978@css.css.sfu.ca> The Fallacy of Equivocation: Boden and Niklasson. In a fairly recent paper (Connection Science, Vol. 12, 2000), Boden and Niklasson purport to demonstrate that a collection of connectionist networks (call them c-nets) can display an important type of Strong Semantic Systematicity. They make frequent references to my 1994 definitions of semantic systematicity and to my papers on this important topic. They also acknowledge that in 1994 I published definite reservations about claims by Niklasson and van Gelder to have produced a connectionist system that displays strong systematicity. In their recent (2000) paper, Boden and Niklasson purport to have answered my reservations by producing a case where a "novel test sentence" is assigned an appropriate meaning representation by previously trained c-nets. Readers may recall that my 1994 definition of strong semantic systematicity required that the "previously trained c-net" must assign an appropriate (and correct) meaning representation to a novel test sentence which contains PREVIOUSLY KNOWN words in at least one novel position. In contrast to this requirement, the putative novel test sentence that Boden and Niklasson employ does not present any previously known words in a novel position. Rather, it presents a purportedly novel word in a known position. However, there is a much more serious problem with their "novel test sentence" (call this sentence S). Here's the problem: The supposed novel sentence S does not produce a correct response when it is first presented to the trained c-net. So, Boden and Niklasson proceed to TRAIN the c-net on the sentence S for an additional 1000 epochs (over and above the earlier training phase). In this latter training phase, only S is presented as input, and backpropagation is employed. Once this further training is complete, Boden and Niklasson contend that a "novel" word in S has now been assigned a meaning representation which they believe to be correct. But, of course, S is no longer a "novel test sentence" at this stage. The c-net has been subjected to intensive training upon S, and only after this further training is complete are Boden and Niklasson able to claim success. Given this, for Boden and Niklasson to describe S as a novel test sentence is (to express the matter diplomatically) to committ a serious instance of the fallacy of equivocation. Indeed, I find it difficult to believe that Boden and Niklasson could be unaware that, as most connectionists use the phrase "test data" (or "novel test sentence"), sentence S is NOT a novel test sentence at all. For this reason, it astonishes me that Boden and Niklasson claim that they have NOW produced an experimental result that satisfactorily answers my 1994 reservations about the results published by Niklasson and van Gelder. My 1994 reservations involved my 1994 definition of strong systematicity, and that definition employed "novel test sentence" in the sense that connectionists commonly employ. At best, Boden and Niklasson are assigning some new, and surprising sense to that phrase -- hence the fallacy of equivocation. I believe there are other serious problems with Boden and Niklasson's (2000) paper, and I am presently writing a detailed critique of that paper. I'll make my new paper available on the internet within a few weeks. Look for a notice of my new critique on "Connectionist List" or send me an email request for the pdf file. In astonishment, Bob Hadley Reference: Boden, M. and Niklasson, L. (2000) "Semantic Systematicity and Context in Connectionist Networks", Connection Science, Vol. 12(2), pp. 111-142. ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Robert F. Hadley (Bob) Phone: 604-291-4488 Professor email: hadley at cs.sfu.ca School of Computing Science and Cognitive Science Program Simon Fraser University Burnaby, B.C. V5A 1S6 Canada Web page: www.cs.sfu.ca/~hadley/ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ From sfr at unipg.it Thu Jul 25 06:36:40 2002 From: sfr at unipg.it (Simone G.O. Fiori (An)) Date: Thu, 25 Jul 2002 12:36:40 +0200 Subject: Mathematical analysis of unsupervised learning systems. Message-ID: <1.5.4.32.20020725103640.01bc3504@unipg.it> Dear Colleagues, I would like to announce the availability of three new papers devoted to the mathematical analysis of unsupervised neural learning systems. Sincerely, Simone Fiori Unsupervised Neural Learning on Lie Group ============================================================================= Author: S. Fiori, Faculty of Engineering, University of Perugia (Italy) Journal: International Journal of Neural Systems Abstract: The present paper aims at introducing the concepts and mathematical details of unsupervised neural learning with orthonormality constrains. The neural structures considered are single non-linear layers and the learnable parameters are organized in matrices, as usual, which gives the parameters spaces the geometrical structure of the Euclidean manifold. The constraint of orthonormality for the connection-matrices further restrict the parameters spaces to differential manifolds such as the orthogonal group, the compact Stiefel manifold and its extensions. For these reasons, the instruments for characterizing and studying the behavior of learning equations for these particular networks are provided by the differential geometry of Lie groups. Although the considered class of learning theories is very general, in the present paper special attention is paid to unsupervised learning paradigms. Download from: http://www.unipg.it/~sfr/publications/IJNS02.ps [46 pages, 565 KB] Notes on Bell-Sejnowski PDF-Matching Neuron ============================================================================= Author: S. Fiori, Faculty of Engineering, University of Perugia (Italy) Journal: Neural Computation Extended abstract: Independent component analysis (ICA) is an emerging neural signal processing technique that allows representing sets of signals as linear combinations of statistically independent bases. In particular, the results of recent investigations about the statistical properties of natural images in relation to the properties of simple cells in V1, suggest that these cells learn to form spatial filters that perform an independent component analysis of the images. One of the most interesting aspects of the ICA theory proposed by Bell and Sejnowski (1996) is the emerging ability of the neurons in the structural model to align to the statistical distributions of the stimuli. Such observation was successfully exploited in order to design different learning rules for blind separation, blind deconvolution and probability density function estimation. In the present paper we consider the basic Bell-Sejnowski class of neuron models and recall the maximum-entropy adapting formulas. By properly selecting a model in the class, that gives rise to tractable mathematics, we are able to present the closed-form expressions of the learning equations, that we particularize for some special excitations. Our main goal is to discuss the features of the neuron-model in an analytical way, in order to gain a deeper insight into the behavior of the equations governing information-theoretic non-linear unit learning. Download from: http://www.unipg.it/~sfr/publications/NeCo2002.zip [8 pages, 576 KB] Information-Theoretic Learning for FAN Network Applied to Eterokurtic Component Analysis ============================================================================= Author: S. Fiori, Faculty of Engineering, University of Perugia (Italy) Journal: IEE Proceedings -- Image, Vision, and Signal Processing Extended abstract: In this paper we deal with instantaneous linear mixtures and focus on the stream of INFOMAX learning algorithms. The paper is devoted to the separation of mixed independent signals from their linear mixtures when the observations are mixed plati-kurtic and lepto-kurtic signals, that is referred to as hybrid or eterokurtic sources problem. We propose the use of networks formed by unsupervised adaptive activation function neurons (FAN), which provide a natural way of estimating the high-order statistical features required to achieve separation. Through numerical and analytical studies the effectiveness of the presented approach is also illustrated and discussed. In Section 2 the problem at hand is formally presented and the adaptive activation function structure is shown to emerge as a natural solution. In Section 3 the general unsupervised learning theory for the FAN neuron is derived, along with a closely-related one, based on a mixture-of-kernel architecture, which is considered for further numerical and architectural comparisons. In Section 4 four different FAN structures are proposed and discussed, while Section 5 is devoted to computer simulations and comparisons. Download from: http://www.unipg.it/~sfr/publications/EKA2002.zip [27 pages, 483 KB] =================================================== Dr Simone Fiori (Mr, EE, PhD)- Assistant Professor Faculty of Engineering - Perugia University Via Pentima bassa, 21 - 05100 TERNI (Italy) eMail: sfr at unipg.it - Fax: +39 0744 492925 Web: http://www.unipg.it/~sfr/ =================================================== From J.A.Bullinaria at cs.bham.ac.uk Fri Jul 26 08:02:47 2002 From: J.A.Bullinaria at cs.bham.ac.uk (John A Bullinaria) Date: Fri, 26 Jul 2002 13:02:47 +0100 (BST) Subject: PhD Studentship/Scholarship Message-ID: =========================================================================== Preliminary Announcement of an Anticipated PhD Studentship/Scholarship ---------------------------------------------------------------------- The School of Computer Science, the University of Birmingham, UK, anticipates (subject to final approval and exchange of contract) setting up a highly competitive PhD Studentship/Scholarship in the Natural Computation group to work on the project, "Automatic Problem Decomposition Using Co-evolution and Ensembles", funded by Honda R&D Europe (Deutschland) GmbH. The PhD Studentship/Scholarship is valued at approximately 17,000 pounds to 20,000 pounds per annum (covering tuition fees, maintenance costs, and travel to the Honda R&D Europe office in Germany) for up to three years (subject to the satisfactory progress of the postholder). In addition, conference travels to present accepted papers may also be funded by Honda R&D Europe or the School of Computer Science on a case by case basis. We are looking for an outstanding candidate to take up this PhD Studentship/Scholarship. The successful candidate must have a first class honours or equivalent in computer science or a closely related field. If your grades are not classified as in a British honours degree, you need to show that you are within at least the top 3% of your year in your average mark. If you have some work/research experience already, please send your best paper with your application. The successful candidate is expected to spend approximately 8 weeks each year at Honda R&D Europe. The School of Computer Science at the University of Birmingham has a strong research group in natural computation, with eight members of academic staff (six faculty and two research fellows) currently specialising in this field: Dr. John Bullinaria (Neural Networks, Evolutionary Computation, Cog.Sci.) Dr. Ke Chen (Neural Networks, Pattern Recognition, Machine Perception) Dr. Aniko Ekart (Genetic Programming, AI, Machine Learning) Dr. Jun He (Evolutionary Computation, Artificial Immune Systems) Dr. Julian Miller (Evolutionary Computation, Machine Learning) Dr. Jon Rowe (Evolutionary Computation, AI) Dr. Thorsten Schnier (Evolutionary Computation, Engineering Design) Prof. Xin Yao (Evolutionary Computation, NNs, Nature Inspired Comp.) Other staff members also working in these areas include Prof. Aaron Sloman (evolvable architectures of mind, co-evolution, interacting niches), Dr. Jeremy Wyatt (evolutionary robotics, classifier systems), and Dr Ela Claridge (evolutionary image processing). There are more than a dozen PhD students currently working in this field. For further information on the technical issues related to this PhD Studentship/Scholarship, please contact Prof. Xin Yao (x.yao at cs.bham.ac.uk). For application and anything else, please contact Dr Peter Hancox, the Research Student Admissions Tutor (p.j.hancox at cs.bham.ac.uk). More information about the PhD programme in the School of Computer Science can be found at http://www.cs.bham.ac.uk/study/postgraduate-research/. ============================================================================ From Gunnar.Raetsch at anu.edu.au Mon Jul 29 03:55:07 2002 From: Gunnar.Raetsch at anu.edu.au (Gunnar Raetsch) Date: Mon, 29 Jul 2002 17:55:07 +1000 Subject: Machine Learning Summer School 2003 Message-ID: <3D44F4DB.20305@anu.edu.au> Machine Learning Summer School The Australian National University, Canberra, Australia 2nd - 15th of February, 2003 ----------------------------------------------------------------------- We would like to inform you that *The Australian National University* will be hosting a Machine Learning Summer School. The School will consist of three courses and a series of special talks and short courses taught by experts from Australia and overseas. The School will be held between February 2 and February 15, 2003. It is suitable for all levels, both for people without previous knowledge in machine learning and those wishing to broaden their expertise in this area. It will allow the participants to get in touch with international experts in this field. The list of courses includes: + Information Geometry, *Shun-Ichi Amari*, RIKEN + Unsupervised Learning, *Zoubin Ghahramani*, Gatsby Institute + Concentration Inequalities, *Gabor Lugosi*, Pompeu Fabra University We are offering a limited number of scholarships (AU$500 and a waiver of the registration cost) for students with a strong academic background. The application deadline is December 1, 2002. Students who are interested should include their CV with their application. Moreover, students may sign up for volunteer work in order to have their registration fees waived. The registration cost of the school is AU$1,200 per person for participants from industry and AU$450 per person for academics. Students are eligible for a further discount and may register for AU$150 per person. All prices are in Australian dollars and include GST. The closing date for early registrations is, December 31, 2001. Registrations received after this date will be subject to 33% surcharge. For further information see our website at http://mlg.anu.edu.au/summer2003 or send e-mail to ml2003 at mlg.anu.edu.au. Regards, Shahar Mendelson, Gunnar Raetsch and Alex Smola -- +-----------------------------------------------------------------+ Gunnar Raetsch http://mlg.anu.edu.au/~raetsch Australian National University mailto:Gunnar.Raetsch at anu.edu.au Research School for Information Tel: (+61) 2 6125-8647 Sciences and Engineering Fax: (+61) 2 6125-8651 Canberra, ACT 0200, Australia From terry at salk.edu Mon Jul 29 18:26:17 2002 From: terry at salk.edu (Terry Sejnowski) Date: Mon, 29 Jul 2002 15:26:17 -0700 (PDT) Subject: NEURAL COMPUTATION 14:9 In-Reply-To: <200206010019.g510Jd343973@purkinje.salk.edu> Message-ID: <200207292226.g6TMQHX88979@purkinje.salk.edu> Neural Computation - Contents - Volume 14, Number 9 - September 1, 2002 ARTICLE Scalable Hybrid Computation with Spikes Rahul Sarpeshkar and Micah O'Halloran NOTES Learning Nonregular Languages: A Comparison of Simple Recurrent Networks and LSTM J. Schmidhuber, F. Gers and D. Eck Center-Crossing Recurrent Neural Networks for the Evolution of Rhythmic Behavior Boonyanit Mathayomchan, Randall D. Beer Reply to Carreira-Perpinan and Goodhill Nicholas V. Swindale, Doron Shoham, Amiram Grinvald, Tobias Bonhoeffer and Mark Hubener LETTERS Dynamics of the Firing Probability of Noisy Integrate-and-Fire Neurons Nicolas Fourcaud and Nicolas Brunel Integrate-and-Fire Neurons Driven by Correlated Stochastic Input Emilio Salinas and Terrence Sejnowski Preintegration Lateral Inhibition Enhances Unsupervised Learning M. W. Spratling and M.H. Johnson On Optimality in Auditory Information Processing Mattias F. Karlsson and John W. C. Robinson Computational Capacity of an Odorant Discriminator: The Linear Separability of Curves N. Caticha, J. E. Palo Tejada, D. Lancet and E. Domany Mixture of Experts Classification Using a Hierarchical Mixture Model Michalis K. Titsias and Aristidis Likas On the Emergence of Rules in Neural Networks Stephen Jose Hanson and Michiro Negishi ----- ON-LINE - http://neco.mitpress.org/ SUBSCRIPTIONS - 2002 - VOLUME 14 - 12 ISSUES USA Canada* Other Countries Student/Retired $60 $64.20 $108 Individual $88 $94.16 $136 Institution $506 $451.42 $554 * includes 7% GST MIT Press Journals, 5 Cambridge Center, Cambridge, MA 02142-9902. Tel: (617) 253-2889 FAX: (617) 577-1545 journals-orders at mit.edu ----- From d.mareschal at bbk.ac.uk Tue Jul 30 07:46:08 2002 From: d.mareschal at bbk.ac.uk (Denis Mareschal) Date: Tue, 30 Jul 2002 12:46:08 +0100 Subject: No subject Message-ID: Dear all, The following recent special issue of Developmental Science may be of interest to readers of this list. It provides an introduction to and overview of current cutting-edge brain imaging methods for use during human infancy. cheers, Denis ======================================================= Contents List Developmental Science 5:3 Special Issue Invited Co-Editors: BJ Casey and Michelle de Haan Title: Imaging Techniques and their Application to Developmental Science 1. Preface: Mark Johnson 2. Introduction: BJ Casey & Michelle de Haan 3. Basic Principles of MRI and Morphometry Studies of Human Brain Development: D. Kennedy, N. Makris, M. R. Herbert, T. Takahashi & V. S. Caviness Jr. 4. Magnetic resonance approaches to the identification of focal pathophysiology in children with brain disease: D. Gadian 5. Monitoring brain development with Quantitative Diffusion Tensor Imaging: A. Ulug 6. Mapping the development of white matter tracts with Diffusion Tensor Imaging: T. Li 7. Functional Magnetic Resonance Imaging : Basic principles of and application to developmental science: BJ Casey, M. Davidson & B. Rosen 8. Application of Pharmacological fMRI with developmental disorders: C. Vaidya 9. Basic principles and applications of ERP/EEG/Intracranial Methods: M. Taylor, P. Sabatier & T. Baldeweg 10. Applications of ERP and fMRI techniques to developmental science: M. de Haan & K. Thomas 11. Functional brain imaging of childhood clinical disorders with PET and SPECT: A. De Volder 12. Magnetoencephalography in pediatric neuroimaging: R. Paetau 13. Basic principles of Optical Imaging and application to the study of infant development: J. Meek 14. Transcranial Magnetic Stimulation in child psychiatry: Disturbed motor system excitability in hypermotoric syndromes: G. H. Moll 15. Commentary: A developmental psychologist looks ahead at developmental neuroimaging research: L. Spelke More details detals about Developmental Science can be obtained form http://www.blackwellpublishers.co.uk/asp/journal.asp?ref=1363-755X ================================================= Dr. Denis Mareschal Centre for Brain and Cognitive Development School of Psychology Birkbeck College University of London Malet St., London WC1E 7HX, UK tel +44 (0)20 7631-6582/6226 reception: 6207 fax +44 (0)20 7631-6312 http://www.psyc.bbk.ac.uk/staff/dm.html ================================================= From smyth at ics.uci.edu Tue Jul 30 14:06:01 2002 From: smyth at ics.uci.edu (Padhraic Smyth) Date: Tue, 30 Jul 2002 11:06:01 -0700 Subject: 2 research positions in machine learning at UC Irvine Message-ID: <3D46D589.F83F985C@ics.uci.edu> We anticipate having 2 postdoc positions in the area of probabilistic learning at UC Irvine for a new NSF-funded project, involving applications to Web and text data. See details below. I would like to encourage readers of this list to apply if interested. Please feel free to distribute to other colleagues. Apologies in advance if you receive multiple copies from cross-posting. Padhraic Smyth Postdoctoral Research Positions in Machine Learning Department of Information and Computer Science University of California, Irvine The Department of Information and Computer Science (ICS) anticipates two full-time research positions available in the area of machine learning and data mining. Responsibilities will include conducting advanced research in predictive and stochastic modeling of large data sets involving events over time (such as Web log data) as well as algorithm development for information extraction and analysis of text streams. Further responsibilities include research demonstrations and presentations, as well as collaborating with graduate students and faculty in the ICS department. Applicants must have earned a Ph.D. in Computer Science, Electrical Engineering, Applied Mathematics, Statistics, or a closely-related discipline, with an emphasis on machine learning or applied statistics. Applicants must have specific knowledge of probabilistic learning methods (such as the EM algorithm and Bayesian learning). Experience with analysis and modeling of large data sets (such as text and Web data) is desirable. The salary range for this position is $45,048 to $54,240 annually, commensurate with training and experience. The initial appointment will be for a twelve-month period with extensions for further years dependent in part on the availability of extra-mural funding. Interested applicants should respond no later than the closing data of September 30th, 2002, by forwarding a cover letter, Curriculum Vitae, and the names of three references to: Professor Padhraic Smyth Department of Information and Computer Science University of California, Irvine CA 92697-3425 or via email to smyth at ics.uci.edu The University of California, Irvine, is an Equal Opportunity Employer, committed to excellence through diversity. From janet at psy.uq.edu.au Tue Jul 30 23:55:52 2002 From: janet at psy.uq.edu.au (Janet Wiles) Date: Wed, 31 Jul 2002 13:55:52 +1000 (EST) Subject: Faculty Position at the University of Queensland Message-ID: I would like to encourage readers of this list interested in computational and cognitive neuroscience to apply for this position. - Janet -------------------------------------------------------------------------- Senior Lecturer/Associate Professor in Neuroscience SCHOOL OF PSYCHOLOGY University of Queensland, Brisbane, Australia The School of Psychology is seeking applications for a continuing Senior Lecturer (Level C) or Associate Professor (Level D) position in any area of neuroscience, although preference may be given to an applicant in the field of cognitive neuroscience, and consideration will be given to the extent to which the applicant's research profile complements the research strengths of the School. The School of Psychology is one of the largest and most prestigious schools of psychology in Australia. It is internationally recognised for research strengths across the breadth of psychology - cognitive psychology, psychophysiology, and clinical neuropsychology are areas of particular research strength. Selection will be based primarily on research standing, teaching excellence, and capacity to attract and supervise postgraduate research students. Applicants should possess a PhD in psychology. At Level C, the successful appointee will have a developing international reputation for his or her research in neuroscience, have a strong research track record in relation to both publication and external research income, a reputation for high quality teaching, and be an experienced research supervisor in psychology. At Level D, a strong international research reputation, an outstanding research track record, a reputation for high quality teaching, and evidence of successful supervision of postgraduate students in psychology are required. The successful appointee will be expected to pursue a strong and productive program of research in neuroscience, to supervise psychology honours and postgraduate research theses in neuroscience, to contribute to the teaching of neuroscience in the undergraduate and honours teaching programs in the School, and to strengthen the links between the School and other centres and schools involved in neuroscience research at the University of Queensland. The remuneration package will be in the range $65,665- $75,716 per annum for Level C plus 17% employer superannuation contributions, and between $79,066- $87,106 per annum for Level D plus 17% employer superannuation contributions. Obtain the position description and selection criteria online or contact Ms Natasha Centis on +61 7 3365 6444 or n.centis at psy.uq.edu.au. Contact Professor Deborah Terry at d.terry at psy.uq.edu.au or telephone +61 7 3365 6220 to discuss the position in more detail. Send applications to the Principal Personnel Officer, Faculty of Social and Behavioural Sciences, University of Queensland, Brisbane, QLD 4072, Australia, or email j.laing at admin.uq.edu.au Closing date for applications: 23 September 2002 Reference Number: 3006530 -- Professor Deborah J. Terry Head of School School of Psychology University of Queensland Brisbane QLD 4072 Australia Phone: +61 7 3365 6220 Fax: +61 7 3365 4466 Email: deborah at psy.uq.edu.au