From jon at incsys.com Mon Apr 1 17:21:12 1991 From: jon at incsys.com (Jon Shultis) Date: Mon, 01 Apr 91 17:21:12 -0500 Subject: Informal Computing Workshop Message-ID: <9104012221.AA12203@incsys.com> Workshop on Informal Computing 29-31 May 1991 Santa Cruz, California Fundamental questions about the nature of informality are gaining importance in computer science. What is informal understanding? What is the nature of informal reasoning? Why is it so powerful and efficient? How are the inconsistency, vagueness, and incompleteness of informal thought managed? How does natural language manage to communicate informal knowledge and reasoning? Computer applications in many fields, ranging from economics and medicine to software engineering and artificial intelligence, demand effective and cognitively accurate answers to these questions in order to capture, represent, and process informal information in computer systems. Inspired by trends toward formalization in logic, mathematics, linguistics, and philosophy, computer scientists historically have tended to regard informal processes as approximate, or imperfect, realizations of formal ideals. Increasingly, however, the idea that informal languages, ontology, and reasoning can (or should) be reduced to (or supplanted by) regimented and "perfected" formalisms is being challenged. Far from being flawed formalisms, informal processes are emerging as fundamental to human understanding and language. From the "informalist" perspective, formalism has been mistaken for the paradigm of intelligence, rather than simply a useful outgrowth of intelligence. The purpose of the Workshop on Informal Computing is to define the study of Informalism, and to begin a coordinated attack on the fundamental issues and problems of the field, bringing together the insights and experience of those who have been working to understand informality in specialized domains. Discussion at the workshop will focus on three major themes: informal knowledge and reasoning; modelling and interpretation; and conversational computing and adaptive languages. Relevant topics include, but are not limited to: intentionality and consciousness; dialogue management; informal meaning and pragmatics; evidential reasoning and belief; resource- and information-limited reasoning; neurocomputation; lessons and techniques from computational linguistics; dynamical and chaotic representations and reasoning; and philosophy of language. The program will be divided between hour-long presentations by invited speakers, and discussion sessions aimed at defining and clarifying informal computing issues, and at identifying promising directions and approaches for future research. The discussion sessions should provide ample opportunity for participants to exchange views, and the schedule will be flexible enough to permit impromptu presentations as appropriate. Also, a follow-up conference may be organized if there is sufficient interest. We are busy making arrangements for speakers and drawing up the schedule, but the basic plan is to devote one day to each of the three themes mentioned above. A preliminary list of speakers includes Bruce d'Ambrosio (Oregon State University) Sandra Carberry (University of Delaware) David Fisher (Incremental Systems) Donald Good (Computational Logic) David Mundie (Incremental Systems) Larry Reeker (IDA) Jeff Rothenberg (RAND) Jon Shultis (Incremental Systems) Tim Standish (University of California at Irvine) Edward Zalta (Stanford University) The final program will be announced on or before 8 May 1991. If you are interested in participating in the workshop, please submit, by 12 April 1991, a brief summary of your interests, and previous or ongoing research that is relevant to the workshop themes. The summaries will be reviewed, and notices of acceptance sent out on 26 April 1991, together with local arrangements information. Summaries should be sent to Jon Shultis Incremental Systems Corporation 319 South Craig Street Pittsburgh, PA 15213 e-mail: jon at incsys.com tel: (412) 621-8888 FAX: (412) 621-0259 Funding for the Workshop on Informal Computing is being provided by DARPA/ISTO in conjunction with ongoing research at Incremental Systems Corporation on adaptive languages for software engineering. From RICKHAAN at ccit.arizona.edu Mon Apr 1 10:50:00 1991 From: RICKHAAN at ccit.arizona.edu (Bigamy is having one wife too many. Monogamy is the same.) Date: Mon, 1 Apr 91 08:50 MST Subject: Announcement: Conference for Philosophy and Psychology Message-ID: <9878EAB1A37F60AC4C@ccit.arizona.edu> PROGRAM, RESERVATION FORMS AND TRAVEL INFORMATION FOR SOCIETY FOR PHILOSOPHY AND PSYCHOLOGY 1991 MEETING SAN FRANCISCO STATE UNIVERSITY JUNE 9-12, 1991 MEETING SITE: Seven Hills Conference Center (west side of the San Francisco State University campus, off Lake Merced Blvd.) PROGRAM N = Nobhill Room P = Presidio Room SUNDAY, JUNE 9 9:00-11:00 Symposium I: Meaning Holism and Conceptual Role Semantics (N) Chair: Janet Levin (Philosophy, USC) Speakers: 1. Ernie Lepore and Jerry Fodor (Philosophy, Rutgers University) 2. Ned Block (Linguistics and Philosophy, MIT) 3. Michael Devitt (Philosophy, U Md/College Park) 11:00-11:15 break 11:15-12:15 Contributed paper 1 (N) Chair: Speaker: Christopher Gauker (Philosophy, University of Cincinnati) "Similarity judgments" Commentator: Robert MacCauley (Philosophy, Emory University) 11:15-12:15 Contributed paper 2 (P) Chair: Jane Duran (Philosophy, UC/Santa Barbara) Speaker: Dan Riesberg (Psychology, Reed College) and Deborah Chambers (Psychology, North Dakota State U.) "Images depict; images describe" Commentator: Kyle Cave (Psychology, Vanderbilt) 12:15-1:15 LUNCH 1:15-2:15 Contributed paper 3 Chair: Larry Birnbaum (Institute for the Learning Sciences, Northwestern University) Speaker: David Kirsh (Cognitive Science, UC/San Diego) "Some problems with the logicist view of artificial intelligence" Commentator: Pat Hayes (MCC) 1:15-2:15 Contributed paper 4 Chair: Irving Krakow (Philosophy, Camden County College) Speaker: Carol Slater (Psychology, Alma College) "Explaining behavior: why we can't lean on learning" Commentator: Joe Levine (Philosophy, NC State/Raleigh) 2:15-2:30 break 2:30-4:30 Symposium II: Is mathematics innate? (N) Chair: Speakers: 1. Karen Wynn (Psychology, U Arizona/Tucson) "Psychological evidence against empiricist theories of mathematical knowledge" 2. Robert Schwartz (Philosophy, U Wisconsin/Milwaukee) "Is mathematics innate?" 3. Philip Kitcher (Philosophy, UCSD) 4:30-4:45 break 4:45-6:00 Keynote address (N) Chair: Jerry Fodor (Philosophy, Rutgers University) Speaker: Steve Pinker (Brain and Cognitive Science, MIT) Title: "Rules and associations in human language" MONDAY, JUNE 10 9:00-11:00 Symposium III: The Roots of Social Cognition (N) Chair: Speakers: 1. Leda Cosmides (Psychology, UC/Santa Barbara) "The logic of social exchange versus the logic of logic" 2. Alan Fiske (Psychology, U Penn) "Innate hypotheses and cultural parameters for social relations" 3. Donald Symons (Anthropology, UC/Santa Barbara) 11:00-11:15 break 11:15-12:15 Contributed paper 5 Chair: Owen Flanagan (Philosophy, Wellesley College) Speaker: Janet Andrews (Psychology, Vassar College) "The role of prototypes in understanding category concepts: a critical assessment" Commentator: Margery Lucas (Psychology, Wellesley College) 11:15-12:15 Contributed paper 6 Chair: Speaker: Anthony Dardis (Philosophy, University of Georgia/Athens) "Content externalism and causal relevance" Commentator: Carol Cleland (Philosophy, University of Colorado/Boulder) 12:15-1:15 LUNCH 1:15-2:15 Contributed paper 7 Chair: Paul Pietroski (Philosophy, McGill Univ.) Speaker: Alvin Goldman (Philosophy, University of Arizona/Tucson) "Mental concepts, self-ascription, and consciousness" Commentator: Paul Boghossian (Philosophy, University of Michigan) 1:15-2:15 Contributed paper 8 Chair: Speaker: Naomi Goldblum (Psychology, Hebrew University) "The distinctiveness of metaphor: a psycholinguistic study" Commentator: Eva Kittay (Philosophy (SUNY/Stony Brook) 2:15-2:30 break 2:30-4:30 Symposium IV: Are Humans Rational? (N) Chair: ?? Speakers 1. Stuart Silvers (Philosophy, Clemson University) "Cognitive performance and assessment, epistemological norms, and the relativity of reason" 2. John Tooby (Psychology, UC/Santa Barbara) "Ecological rationality" 3. Edward Stein (Philosophy, Williams College & MIT) 7:00-8:30 Banquet (N) 8:30-10:00 Presidential Address (N) Chair: Patricia Kitcher (Philosophy, UCSD) Speaker: Ray Jackendoff (Program in Linguistics and Cognitive Science, Brandeis) Title: "Is there a faculty of social cognition?" TUESDAY, JUNE 11 9:00-10:00 Invited talk (N) Chair: Kent Bach (Philosophy, San Francisco State) Speaker: Michael Scriven (Pacific Graduate School of Psychology, Palo Alto, CA) Title: "Implicit rules in logic" 10:00-10:15 break 10:15-12:15 Symposium V: Prototype Theory and Concept Stability Chair: Ken Livingston (Psychology, Vassar College) Speakers 1. Georges Rey (Philosophy, U Md/College Park) "Concepts and reference fixers" 2. Alison Gopnik (Psychology, UC/Berkeley) "Concepts as theories: who's afraid of semantic holism" 3. George Bealer (Philosophy, UColorado/Boulder) "Philosophical constraints on a theory of concepts" 12:15-1:15 LUNCH 1:15-2:15 Contributed papers 9 (N)+(P) Chair: A.A. Howsepian (Philosophy, Notre Dame) Speakers: G. Lynn Stephens & George Graham (Philosophy, Univ of Alabama/Birmingham) "Introspective identification and disturbances of self consciousness" Commentator: Jennifer Church (Philosophy, Vassar College) 1:15-2:15 Contributed papers 10 (N)+(P) Chair: Anti Bax (Clemson University) Speaker: Kirk Ludwig (Philosophy, University of Florida/Gainesville) "How can reasons be causally relevant to what we do?" Commentator: Fred Dretske (Philosophy, Stanford University) 2:15-2:30 break 2:30-3:30 Contributed papers 11 (N)+(P) Chair: Speaker: Louise Antony (Philosophy, NC State University/Raleigh) "Rabbit pots and supernovas: on the relevance of psychological data to linguistic theory" Commentator: Scott Soames (Philosophy, Princeton University) 2:30-3:30 Contributed paper 12 (N)+(P) Chair: Speaker: Peter Ludlow (Philosophy, SUNY/Stony Brook) "In defense of the dual aspect theory" Commentator: Barbara Von Eckardt (University of Nebraska/Lincoln) 3:30-3:45 break 3:45-5:45 Contributed papers 13 & 14 (P) [B Chair: Jeffrey Poland (Psychology, U Nebraska/Lincoln) Speaker: Rob Wilson (Philosophy, Cornell University) "Individualism and causal powers" Commentator: Terry Horgan (Philosophy, Memphis State University) Speaker: Joseph Owens (Philosophy, University of Minnesota/Twin Cities) "Psychophysical supervenience and inner access" Commentator: Piers Rawling (Philosophy, University of Georgia/Athens) 3:45-5:45 Invited lecture (N) Chair: Speaker: Randy Gallistel (Psychology, UC/Berkeley) Title: "Is the associative bond the phlogiston of psychology?" Discussants: 1. Paul Churchland (Philosophy, UCSD) 2. Ben Williams (Psychology, UCSD) 5:45-6:30 Business meeting (N) WED JUNE 12 9:00-12:00 Symposium VI: Consciousness Chair: Speakers: 1. Dan Dennett (Philosophy, Tufts University) "Time and the brain: escape from the theater of consciousness" 2. Robert Van Gulick (Philosophy, Syracuse University) "Understanding the phenomenal mind" 3. Benjamin Libet (Physiology, School of Medicine, UC/San Francisco) "The cerebral time-on theory for conscious and unconscious mental function" 4. David Rosenthal (Philosophy, Graduate Center/CUNY) ........................................ Accomodations: Lodging is available at reasonable rates at two on campus locations. Both the Guest Center (hotel-like) and the Residence Halls (shared bath) are very near the Conference Center. Accomodations are available only in a complete package which includes 4 nights lodging (begining night of June 8) and breakfast and lunch each day of the conference. Rates are given on the Reservation Form (below). Limited availability for the Guest Center, so reservations will be taken in order received. (If you request the Guest Center and end up booked in the Residence Hall, the difference will be refunded at registration.) Please note that the SPP must supply SFSU with a guest list and full payment 30 days in advance to hold reservations. So please send your reservation and payment ASAP and by April 30 at the latest. Doubles can not be reserved until both persons register. STAYING ON: If you want to stay in SF for a few extra days, space should be available. Please indicate such an interest on the registration form. ARRIVAL: Check in time: after 3:00 PM Sat. June 8 (On Saturday 6/8 only - come to lounge of Verducci Hall, the only SFSU building on Lake Merced Boulevard. On other days registration will be at Conference Center) Check out time: 2:00 PM Wed. June 12 CHILDREN: Unfortunately insurance provisions do not permit children in SFSU lodging. .................................................... REGISTRATION & ROOM RESERVATIONS Room and Board (4 nights + 4 breakfasts & lunches) Guest Center: Single: $290. *Double Occ: $210. per person Residence Hall Single: $180. *Double Occ: $120. per person Total: $_____ (*If Double Occupancy, list name of person you are sharing with ___________________________) .................................................... CONFERENCE REGISTRATION: Member: $20 Nonmember: $30 Student: $10. $_________ .................................................... 1991 SPP MEMBERSHIP DUES: Regular: $15 Student: $ 5 (New members may pay 1991 dues and register as members.) $_________ ................................................... PRESIDENTIAL ADDRESS BANQUET June 10 Price: $18.00 $_________ .................................................... PARKING: $6.00 $_______ _ GRAND TOTAL $_________ Your Name______________________________ Address________________________________ MAKE CHECK PAYABLE TO SOCIETY FOR PHILOSOPHY AND PSYCHOLOGY (WILL NOT BE CASHED BEFORE MAY 1) AND MAIL ASAP (AND NO LATER THAN APRIL 30) TO: KENT BACH 66 ST. GERMAIN SAN FRANCISCO, CA 94114 Check here if you wish to know about the possibilities of staying on for extra days and indicate dates _________ GETTING TO SAN FRANCISCO STATE UNIVERSITY BY AIR: Fly to San Franciso International Airport. Shuttles are readily available outside the arrivals area. Look for vans marked SF and just ask the driver if his/her van goes to San Francisco State University. BY CAR: From the North: Take Highway 1 south to 19th and Holloway Avenues. From the South: Take Interstate 280 north; exit at 19th Ave. Take Junipero Serra to Holloway Ave; turn left onto Holloway. From the East: Take Interstate 80 across the Oakland Bay Bridge to Route 101 south to Interstate 280. Take 19th Ave. exit; bear right onto Sagamore Blvd. to Brotherhood Way to Junipero Serra Blvd north. Take Junipero Serra to Holloway Ave., turn left on Holloway to 19th Ave. LOCAL TRANSIT: From downtown San Francisco, take the Muni No. 17 Express (2nd and Market Streets) or the M Streetcar. For further information about any aspect of local arrangements call Kent Bach at 415-665-1040 The Program Chairs wish to thank the following for their help in refereeing contributed papers: Louise Antony, Kent Bach, Ned Block, Denise Cummins, A.J. Figuerado, Owen Flanagan, Alvin Goldman, George Graham, Kerry Green, Eli Hirsch, Terry Horgan, Eva Kittay, Ernie LePore, Dan Lloyd, Chris Maloney, Charles Marks, Bob Matthews, Lynn Nadel, Janet Nicol, Sarah Patterson, Paul Pietroski, William Rappoport, Georges Rey, Mark Richard, Bob Schwartz, Stuart Silvers, Varda Solomon, Ed Stein, Bob Van Gulik, Barbara Von Eckhardt, Stephen White, David Wong, Karen Wynn, Palle Yourgrau Program Chairs: Jerry Samet (Philosophy, Brandeis University) Paul Bloom (Psychology, University of Arizona) E-mail BLOOM at ccit.arizona.edu From ps_coltheart at vaxa.mqcc.mq.oz.au Wed Apr 3 08:44:23 1991 From: ps_coltheart at vaxa.mqcc.mq.oz.au (Max Coltheart) Date: Wed, 3 Apr 91 08:44:23 est Subject: Full Professor Position in Australia Message-ID: <9104022244.AA00985@macuni.mqcc.mq.oz.au> PROFESSOR OF PSYCHOLOGY VACANCY Applications are invited for appointment to the position of Professor of Psychology at Macquarie University in Sydney, NSW, Australia. Applicants with interests in any area of psychology will be considered. Professor M. Coltheart and Professor J. K. Collins currently hold chairs within the department. Further information about the University, conditions of appointment and method of application may be obtained from the Academic Staff Office, Macquarie University, NSW 2109 or by telephoning (02) 805-7390. Closing date for applications 14 June 1991. Initial enquiries may be directed to Professor Max Coltheart (ps_coltheart at vaxa.mqcc.mq.oz.au, telephone (02) 805-8086, confidential fax (02) 805-8127) or Professor John Collins (tel (02) 805-8030) or Associate Professor George Cooney (ps_cooney at vaxa.mqcc.mq.oz.au, telephone (02) 805-8067). From pratt at paul.rutgers.edu Fri Apr 5 17:08:52 1991 From: pratt at paul.rutgers.edu (Lorien Y. Pratt) Date: Fri, 5 Apr 91 17:08:52 EST Subject: Two papers on information transfer / problem decomposition Message-ID: <9104052208.AA08551@paul.rutgers.edu> The following two papers are now available via FTP from the neuroprose archives. The first is for AAAI91, so written towards an AI/Machine learning audience. The second is for IJCNN91, so more neural network-oriented. There is some overlap between them: the AAAI paper reports briefly on the study describved in more detail in the IJCNN paper. Instructions for retrieval are at the end of this message. --Lori #@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@ Direct Transfer of Learned Information Among Neural Networks To appear: Proceedings of AAAI-91 Lorien Y. Pratt and Jack Mostow and Candace A. Kamm Abstract A touted advantage of symbolic representations is the ease of transferring learned information from one intelligent agent to another. This paper investigates an analogous problem: how to use information from one neural network to help a second network learn a related task. Rather than translate such information into symbolic form (in which it may not be readily expressible), we investigate the direct transfer of information encoded as weights. Here, we focus on how transfer can be used to address the important problem of improving neural network learning speed. First we present an exploratory study of the somewhat surprising effects of pre-setting network weights on subsequent learning. Guided by hypotheses from this study, we sped up back-propagation learning for two speech recognition tasks. By transferring weights from smaller networks trained on subtasks, we achieved speedups of up to an order of magnitude compared with training starting with random weights, even taking into account the time to train the smaller networks. We include results on how transfer scales to a large phoneme recognition problem. @%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@ Improving a Phoneme Classification Neural Network through Problem Decomposition To appear: Proceedings of IJCNN-91 L. Y. Pratt and C. A. Kamm Abstract In the study of neural networks, it is important to determine whether techniques that have been validated on smaller experimental tasks can be scaled to larger real-world problems. In this paper we discuss how a methodology called {\em problem decomposition} can be applied to AP-net, a neural network for mapping acoustic spectra to phoneme classes. The network's task is to recognize phonemes from a large corpus of multiple-speaker, continuously-spoken sentences. We review previous AP-net systems and present results from a decomposition study in which smaller networks trained to recognize subsets of phonemes are combined into a larger network for the full signal-to-phoneme mapping task. We show that, by using this problem decomposition methodology, comparable performance can be obtained in significantly fewer arithmetic operations. ^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^% To retrieve: unix> ftp cheops.cis.ohio-state.edu (or 128.146.8.62) Name: anonymous Password: neuron ftp> cd pub/neuroprose ftp> binary ftp> get pratt.aaai91.ps.Z ftp> get pratt.ijcnn91.ps.Z ftp> quit unix> uncompress pratt.aaai91.ps.Z pratt.ijcnn91.ps.Z unix> lpr pratt.aaai91.ps pratt.ijcnn91.ps From ernst at acquine.cns.caltech.edu Fri Apr 5 23:11:43 1991 From: ernst at acquine.cns.caltech.edu (Ernst Niebur) Date: Fri, 5 Apr 91 20:11:43 PST Subject: Connectionist Faculty Position at Boulder In-Reply-To: Paul Smolensky's message of Tue, 23 Oct 90 10:04:58 -0600 <9010231604.AA06346@axon.Colorado.EDU> Message-ID: <9104060411.AA02276@cns.caltech.edu> Please note my new internet address (ernst at cns.caltech.edu instead of the former ernst at descartes.cns.caltech.edu). Also, in the course of changing our system, some messages were lost. If you sent me a message between March 18 and April 5 which I did not answer, it means the message was lost. Please send it again. Sorry for the inconvenience. --Ernst Niebur ========================================================================= Ernst Niebur Phone (818)356-6885 Computation and Neural Systems Fax (818)796-8876 Caltech 216-76 ernst at cns.caltech.edu Pasadena, CA 91125 ernst at caltech.bitnet USA ========================================================================= From ibm at dit.upm.es Mon Apr 8 07:08:01 1991 From: ibm at dit.upm.es (Ignacio Bellido Montes) Date: Mon, 8 Apr 91 13:08:01 +0200 Subject: Neural Nets Workshop (IWANN 91) Message-ID: <9104081108.AA01055@bosco.dit.upm.es> There will be a Neural Nets Workshop Granada (Spain) next September. I think this is the first international workshop of this kind that will be held in Spain, I hope all of you can send contributions and participate. Here I send you the electronic version of the CFP followed by the LaTeX version. If you have any question about the workshop, please ask me or the people of the workshop secretary. I'm not on the organization but I can help finding people etc... About Granada, I can tell you this is one of the most wonderful cities in the Spain and ... in the World. It has one of the most famous and wonder arab contributions to the culture, "La Alhambra", and small and quiet lanes on the old side of the city, the "Albaicin", that transmit the visitor a very good feeling. I hope you be able to participate and enjoy... See you in Granada. Gregorio Fernandez Dpt. Ingenieria de Sistemas Telematicos ETSI Telecomunicacion- UPM Ciudad Universitaria 28040 Madrid, Spain E-mail: gfernandez at dit.upm.es ------------------------------------------------------------------------------- INTERNATIONAL WORKSHOP ON ARTIFICIAL NEURAL NETWORKS IWANN'91 First Announcement and Call for Papers Granada, Spain September 17-19, 1991 ORGANISED AND SPONSORED BY Spanish Chapter of the Computer Society of the IEEE, AEIA (IEEE Affiliate Society), and Department of Electronic and Computer Technology. University of Granada. Spain. SCOPE Artificial Neural Networks (ANN) were first developed as structural or functional modelling systems of natural ones, featuring the ability to perform problem-solving tasks. They can be thought as computing arrays consisting of series of repetitive uniform processors (neuron-like elements) placed on a grid. Learning is achieved by changing the interconnections between these processing elements. Hence, these systems are also called connectionist models. ANN has become a subject of wide-spread interest: they offer an odd scheme-based programming standpoint and exhibit higher computing speeds than conventional von-Neumann architectures, thus easing or even enabling handling complex task such as artificial vision, speech recognition, information recovery in noisy environments or general pattern recognition. In ANN systems, collective information management is achieved by means of parallel operation of neuron-like elements, into which information processing is distributed. It is intended to exploit this highly parallel processing capability as far as possible in complex problem-solving tasks. Cross-fertilization between the domains of artificial and real neural nets is desirable. The more genuine problems of biological computation and information processing in the nervous system still remain open and contributions in this line are more than welcome. Methodology, theoretical frames, structural and organizational principles in neuroscience, self-organizing and co- operative processes and knowledge based descriptions of neural tissue are relevant topics to bridge the gap between the artificial and natural perspectives. The workshop intends to serve as a meeting place for engineers and scientists working in this area, so that present contacts and relationships can be further increased. The workshop will comprise two complementary activities: . scientific and technical conferences, and . scientific communications sessions. TOPICS The workshop is open to all aspects of artificial neural networks, including: 1. Neural network theories. Neural models. 2. Biological perspectives 3. Neural network architectures and algorithms. 4. Software developments and tools. 5. Hardware implementations 6. Applications. LOCATION Facultad de Ciencias Campus Universitario de Fuentenueva Universidad de Granada 18071 GRANADA. (SPAIN) LANGUAGES English and Spanish will be the official working languages. English is preferable as the working language. CALL FOR PAPERS The Programme Committee seeks original papers on the six above mentioned areas. Survey papers on the various available approaches or particular application domains are also sought. In their submitted papers, authors should pay particular attention to explaining the theoretical and technical choices involved, to make clear the limitations encountered and to describe the current state of development of their work. INSTRUCTIONS TO AUTHORS Three copies of submitted papers (not exceeding 8 pages in 21x29.7 cms (DIN-A4), with 1,6 cm. left, right, top and bottom margins) should be received by the Programme Chairman at the address below before June 20, 1991. The headlines should be centered and include: . the title of paper in capitals . the name(s) of author(s) . the address(es) of author(s), and . a 10 line abstract. Three blank lines should be left between each of the above items, and four between the headlines and the body of the paper, written in English, single-spaced and not exceeding the 8 pages limit. All papers received will be refereed by the Programme Committee. The Committee will communicate their decision to the authors on July 10. Accepted papers will be published in the proceedings to be distributed to workshop participants. In addition to the paper, one sheet should be attached including the following information: . the title of the paper, . the name(s) of author(s), . a list of five keywords, . a reference to which of the six topics the paper concerns, and . postal address of one of the authors, with phone and fax numbers, and E-mail (if available). We intend to get in touch with various international publishers (such as Springer-Verlag and Prentice-Hall) for the final version of the proceedings. Contributions to be sent to: Prof. Jose Mira Dpto. Informatica y Automatica UNED C/Senda del Rey s/n 28040 MADRID (Spain) Tel. (34) 1 5 44 60 00 Fax: (34) 1 5 44 67 37 ACCOMMODATION A list of available Hotels will be sent on registration. Hotel reservations can be made directly by each participant with the local agency below. All request should be addressed to: Viajes Internacional Expreso (VIE) Galerias Preciados Carrera del Genil, s/n 18005 GRANADA (Spain) Tel. (34) 58-22.44.95, -22.75.86, -224944 Telex: 78525 We can only guarantee to accept reservation received by July 25. REGISTRATION FEE . Regular fee: 35.000 ptas. . IEEE, AEIA and ATI members fee: 30.000 ptas. . Scholarship holders fee: 5.000 ptas. Inscription payments: Transfer to: IWANN'91 account number: 16.142.512 Caja Postal (Code: 2088-2037.1) Camino de Ronda, 138 18003 GRANADA (SPAIN) or alternatively, cheque made out to: IWANN'91 (16.142.512) Secretariat address: Departamento de Electronica y Tecnologia de Computadores Facultad de Ciencias Universidad de Granada 18071 GRANADA (SPAIN) FAX: 34-58-24.32.30 or 34-58-27.42.58 Phone: 34-58-24.32.26 E-Mail: jmerelo at ugr.es aprieto at ugr.es PROGRAM AND ORGANIZATION COMMITTEE Senen Barro Univ. de Santiago Joan Cabestany Univ. Pltca. de Catalunya Jose Antonio Corrales Univ. Oviedo. Gregorio Fernandez Univ. Pltca. de Madrid J. Simoes da Fonseca Univ. de Lisboa Antonio Lloris Univ. Granada Javier Lopez Aligue Univ. de Extremadura. Jose Mira (Programme Chairman) UNED. Madrid Roberto Moreno Univ Las Palmas Gran Canaria Alberto Prieto (Organization Chairman) Univ. Granada Francisco Sandoval Univ. de Malaga Carmen Torras Instituto de Cibernetica. CSIC. Barcelona Elena Valderrama CNM- Univ. Autonoma de Barcelona LOCAL ORGANIZING COMMITTEE (Universidad de Granada) Juan Julian Merelo Julio Ortega Francisco J. Pelayo Begona del Pino Alberto Prieto (To be completed and returned as soon as possible to: Departamento de Electronica. Facultad de Ciencias. Univ. de Granada. 18071 GRANADA (SPAIN). FAX (34)-58-24.32.30) Block letters, please Name: Company/Organization: Address: State/Country: E-mail: Phone: Fax: Please tick as appropriate: I intend to: attend to the workshop submit a paper Name(s) of Author(s): Provisional Title: ------------------------------------------------------------------------------ %%Latex-version \documentstyle{article} \setlength{\textheight}{24.0cm} \setlength{\textwidth}{16.0cm} \setlength{\voffset}{-1.0in} \setlength{\hoffset}{-1.0in} \begin{document} \thispagestyle{empty} \vspace*{3cm} \begin{center} {\Large \bf INTERNATIONAL WORKSHOP\\ ON\\ ARTIFICIAL NEURAL NETWORKS\\ \vspace{2cm} IWANN'91\\} {\large \vspace{4cm} First Announcement and Call for Papers\\ \vspace{4cm} Granada, Spain\\ September 17-19, 1991\\ \vspace{4cm} ORGANISED AND SPONSORED BY:\\ Spanish Chapter of the Computer Society of the IEEE,\\ AEIA (IEEE Affiliate Society), and\\ Department of Electronic and Computer Technology\\ (University of Granada, Spain)} \end{center} \newpage \section*{SCOPE} Artificial Neural Networks (ANN) were first developed as structural or functional modelling systems of natural ones, featuring the ability to perform problem-solving tasks. They can be thought as computing arrays consisting of series of repetitive uniform processors (neuron-like elements) placed on a grid. Learning is achieved by changing the interconnections between these processing elements. Hence, these systems are also called connectionist models. ANN has become a subject of wide-spread interest: they offer an odd scheme-based programming standpoint and exhibit higher computing speeds than conventional von-Neumann architectures, thus easing or even enabling handling complex task such as artificial vision, speech recognition, information recovery in noisy environments or general pattern recognition. In ANN systems, collective information management is achieved by means of parallel operation of neuron-like elements, into which information processing is distributed. It is intended to exploit this highly parallel processing capability as far as possible in complex problem-solving tasks. Cross-fertilization between the domains of artificial and real neural nets is desirable. The more genuine problems of biological computation and information processing in the nervous system still remain open and contributions in this line are more than welcome. Methodology, theoretical frames, structural and organizational principles in neuroscience, self-organizing and co- operative processes and knowledge based descriptions of neural tissue are relevant topics to bridge the gap between the artificial and natural perspectives. The workshop intends to serve as a meeting place for engineers and scientists working in this area, so that present contacts and relationships can be further increased. The workshop will comprise two complementary activities: \begin{itemize} \item scientific and technical conferences, and \item scientific communications sessions. \end{itemize} \section*{TOPICS} The workshop is open to all aspects of artificial neural networks, including: \begin{enumerate} \item Neural network theories. Neural models. \item Biological perspectives \item Neural network architectures and algorithms. \item Software developments and tools. \item Hardware implementations \item Applications. \end{enumerate} \section*{LOCATION} Facultad de Ciencias\\ Campus Universitario de Fuentenueva\\ Universidad de Granada\\ 18071 GRANADA (SPAIN)\\ \section*{LANGUAGES} English and Spanish will be the official working languages. English is preferable as the working language. \newpage \section*{CALL FOR PAPERS} The Programme Committee seeks original papers on the six above mentioned areas. Survey papers on the various available approaches or particular application domains are also sought. In their submitted papers, authors should pay particular attention to explaining the theoretical and technical choices involved, to make clear the limitations encountered and to describe the current state of development of their work. \section*{INSTRUCTIONS TO AUTHORS} Three copies of submitted papers (not exceeding 8 pages in 21x29.7 cms (DIN-A4), with 1,6 cm. left, right, top and bottom margins) should be received by the Programme Chairman at the address below before June 20, 1991. The headlines should be centered and include:\\ $\bullet$ the title of paper in capitals\\ $\bullet$ the name(s) of author(s)\\ $\bullet$ the address(es) of author(s), and\\ $\bullet$ a 10 line abstract. Three blank lines should be left between each of the above items, and four between the headlines and the body of the paper, written in English, single-spaced and not exceeding the 8 pages limit. All papers received will be refereed by the Programme Committee. The Committee will communicate their decision to the authors on July 10. Accepted papers will be published in the proceedings to be distributed to workshop participants. In addition to the paper, one sheet should be attached including the following information:\\ $\bullet$ the title of the paper,\\ $\bullet$ the name(s) of author(s),\\ $\bullet$ a list of five keywords,\\ $\bullet$ a reference to which of the six topics the paper concerns, and\\ $\bullet$ postal address of one of the authors, with phone and fax numbers, and E-mail (if available). \vspace{0.5cm} We intend to get in touch with various international publishers (such as Springer-Verlag and Prentice-Hall) for the final version of the proceedings. \vspace{0.5cm} \noindent {\bf Contributions to be sent to:} \noindent Prof. Jos\'{e} Mira\\ UNED---Dpto. Inform\'{a}tica y Autom\'{a}tica\\ C/Senda del Rey s/n\\ 28040 MADRID (Spain)\\ Tel. (34) 1 5 44 60 00; Fax: (34) 1 5 44 67 37\\ \section*{ACCOMMODATION} A list of available Hotels will be sent on registration. Hotel reservations can be made directly by each participant with the local agency below. All request should be addressed to:\\ \noindent Viajes Internacional Expreso (VIE)\\ Galerias Preciados\\ Carrera del Genil, s/n\\ 18005 GRANADA (Spain)\\ Tel. (34) 58-22.44.95, -22.75.86, -224944\\ Telex: 78525\\ We can only guarantee to accept reservation received by July 25. \newpage \section*{REGISTRATION FEE} \begin{tabular}{lr} Regular fee: & 35.000 ptas.\\ IEEE, AEIA and ATI members fee: & 30.000 ptas.\\ Scholarship holders fee: & 5.000 ptas.\\ \end{tabular} \subsection*{Inscription payments: } \begin{tabular}{ll} Transfer to: & IWANN'91\\ & account number: 16.142.512\\ & Caja Postal (Code: 2088-2037.1)\\ & Camino de Ronda, 138\\ & 18003 GRANADA (SPAIN)\\ or alternatively, cheque made out to: & IWANN'91 (16.142.512)\\ \end{tabular} \subsection*{Secretariat address:} Departamento de Electr\'{o}nica y Tecnolog\'{i}a de Computadores\\ Facultad de Ciencias\\ Universidad de Granada\\ 18071 GRANADA (SPAIN)\\ FAX: 34-58-24.32.30 or 34-58-27.42.58\\ Phone: 34-58-24.32.26\\ E-Mail: jmerelo at ugr.es, or: aprieto at ugr.es \section*{PROGRAM AND ORGANIZATION COMMITTEE} \begin{tabular}{lr} Sen\'{e}n Barro & Univ. de Santiago\\ Joan Cabestany & Univ. Pltca. de Catalunya\\ Jose Antonio Corrales& Univ. Oviedo\\ Gregorio Fern\'{a}ndez & Univ. Pltca. de Madrid\\ J. Simoes da Fonseca & Univ. de Lisboa\\ Antonio Lloris & Univ. Granada\\ Javier L\'{o}pez Aligu\'{e} & Univ. de Extremadura.\\ Jos\'{e} Mira {\bf (Programme Chairman)} & UNED. Madrid\\ Roberto Moreno & Univ. Las Palmas de Gran Canaria\\ Alberto Prieto {\bf (Organization Chairman)}& Univ. Granada\\ Francisco Sandoval & Univ. de M\'{a}laga\\ Carme Torras & Instituto de Cibern\'{e}tica. CSIC. Barcelona\\ Elena Valderrama & CNM--Univ. Aut\'{o}noma de Barcelona\\ \end{tabular} \section*{LOCAL ORGANIZING COMMITTEE} {\large \bf(Universidad de Granada)}\\ Juan Juli\'{a}n Merelo\\ Julio Ortega\\ Francisco J. Pelayo\\ Bego\~{n}a del Pino\\ Alberto Prieto \\ \newpage \noindent (To be completed and returned as soon as possible to:\\ Departamento de Electronica, Facultad de Ciencias,\\ Univ. de Granada, 18071 GRANADA, SPAIN,\\ FAX (34)-58-24.32.30)\\ Block letters, please\\ \vspace{0.5cm} Name:\\ Company/Organization:\\ Address:\\ \vspace{1.0cm} State/Country:\\ E-mail: \hspace{4cm} Phone: \hspace{4cm} Fax:\\ Please tick as appropriate:\\ I intend to:\\ $\Box$ attend to the workshop\\ $\Box$ submit a paper\\ Name(s) of Author(s):\\ Provisional Title: \end{document} From SAYEGH at CVAX.IPFW.INDIANA.EDU Wed Apr 10 20:41:22 1991 From: SAYEGH at CVAX.IPFW.INDIANA.EDU (SAYEGH@CVAX.IPFW.INDIANA.EDU) Date: Wed, 10 Apr 1991 20:41:22 EDT Subject: Proceedings of Third NN and PDP Conference, Indiana-Purdue Message-ID: <910410204122.20204a59@CVAX.IPFW.INDIANA.EDU> The proceedings of the THIRD conference on Neural Networks and Parallel Distributed Processing held in April 1990 at Indiana-Purdue University in Ft Wayne can be obtained by writing to: Ms. Sandra Fisher Physics Department Indiana University-Purdue University Ft Wayne, IN 46805 and including $5 + $1 for mailing and handling. Checks should be made payable to The Indiana-Purdue Foundation. The 109 page proceedings contain the following articles: INTEGRATED AUTONOMOUS NAVIGATION BY ADAPTIVE NEURAL NETWORKS Dean A. Pomerleau Department of Computer Science Carnegie Mellon University APPLYING A HOPFIELD-STYLE NETWORK TO DEGRADED PRINTED TEXT RESTORATION Arun Jagota Department of Computer Science State University of New York at Buffalo RECENT STUDIES WITH PARALLEL, SELF-ORGANIZ- ING, HIERARCHICAL NEURAL NETWORKS O.K. Ersoy & D. Hong School of Electrical Engineering Purdue University INEQUALITIES, PERCEPTRONS AND ROBOTIC PATH- PLANNING Samir I. Sayegh Department of Physics Indiana University-Purdue University GENETIC ALGORITHMS FOR FEATURE SELECTION FOR COUNTERPROPAGATION NETWORKS F.Z. Brill & W.N. Martin Department of Computer Science University of Virginia MULTI-SCALE VISION-BASED NAVIGATION ON DIS- TRIBUTED-MEMORY MIMD COMPUTERS A.W. Ho & G.C. Fox Caltech Concurrent Computation Program California Institute of Technology A NEURAL NETWORK WHICH ENABLES SPECIFICATION OF PRODUCTION RULES N. Liu & K.J. Cios The University of Toledo PIECE-WISE LINEAR ESTIMATION OF MECHANICAL PROPERTIES OF MATERIALS WITH NEURAL NETWORKS I.H. Shin, K.J. Cios, A. Vary* & H.E. Kautz* The University of Toledo & NASA Lewis Re- search Center* INFLUENCE OF THE COLUMN STRUCTURE ON INTRA- CORTICAL LONG RANGE INTERACTIONS E. Niebur & F. Worgotter California Institute of Technology LEARNING BY GRADIENT DESCENT IN FUNCTION SPACE Ganesh Mani University of Wisconsin-Madison REAL TIME DYNAMIC RECOGNITION OF SPATIAL TEMPORAL PATTERNS M. F. Tenorio School of Electrical Engineering Purdue University A NEURAL ARCHITECTURE FOR COGNITIVE MAPS Martin Sonntag Cognitive Science & Machine Intelligence Lab University of Michigan P.S. The Fourth Conference is scheduled to start April 11, 91 at 6pm in the Classroom Medical Building, CM159, of the Fort Wayne Campus of Indiana and Purdue. A previous announcement of this conference was made on the list. From andreas%psych at Forsythe.Stanford.EDU Thu Apr 11 00:48:16 1991 From: andreas%psych at Forsythe.Stanford.EDU (Andreas Weigend) Date: Wed, 10 Apr 91 21:48:16 PDT Subject: No subject Message-ID: <9104110448.AA07648@psych> ~t Subject: prepint by Lumer & Huberman: "Binding Hierarchies:..." The following preprint is available in hardcopy form. It can be obtained by sending e-mail to: lumer at parc.xerox.com Please do NOT reply to this message. ______________________________________________________________________ Binding Hierarchies: A Basis for Dynamic Perceptual Grouping E. Lumer and B. A. Huberman Stanford University and Xerox PARC Abstract Since it has been suggested that the brain binds its fragmentary representations of perceptual events via phase-locking of stimulated neuron oscillators, it is important to determine how extended synchronization can occur in a clustered organization of cells possessing a distribution of firing rates. In order to answer that question, we establish the basic conditions for the existence of a binding mechanism based on phase-locked oscillations. In addition, we present a simple hierarchical architecture of feedback units which not only induces robust synchronization within and segregation between perceptual groups, but also serves as a generic binding machine. ______________________________________________________________________ From johni at ee.ubc.ca Wed Apr 10 15:25:53 1991 From: johni at ee.ubc.ca (johni@ee.ubc.ca) Date: Wed, 10 Apr 91 12:25:53 PDT Subject: NIPS 91 information request Message-ID: <9104101925.AA22205@snickers.ee.ubc.ca> Hi, does anybody have information on NIPS '91? I like to know the time, place, contact, dealine for submission of papers, ... Thanks. From leow at mcc.com Thu Apr 11 12:37:53 1991 From: leow at mcc.com (Wee Kheng Leow) Date: Thu, 11 Apr 91 11:37:53 CDT Subject: sharing hotel room Message-ID: <9104111637.AA23020@koala.aca.mcc.com> I'm attending the course and conference on neural networks (Neural Networks: From Foundations to Applications, and Neural Networks for Vision and Image Processing) to be held in Tyngsboro, MA on May 5-12. I've reserved a double room in Red Roof Inn (Nashua, NH) and like to look for a partner to share the room. Interested participant please email to leow at mcc.com (DO NOT reply to this message). Wee Kheng Leow p.s. I'm a male Chinese graduate student and I speak English and Mandarin. From collins at z.ils.nwu.edu Thu Apr 11 12:28:13 1991 From: collins at z.ils.nwu.edu (Gregg Collins) Date: Thu, 11 Apr 91 11:28:13 CDT Subject: Conference announcement: ML91 Message-ID: <9104111628.AA03045@z.ils.nwu.edu> ML91 -- The Eighth International Workshop on Machine Learning Registration Information On behalf of the organizing committee, and the individual workshop committees, we are pleased to announce that ML91, the Eighth International Workshop on Machine Learning, will be held at Northwestern University, Evanston, Illinois, USA, June 27-29, 1991. ML91 comprises eight workshop tracks: o Automated Knowledge Acquisition o Computational Models of Human Learning o Constructive Induction o Learning from Theory and Data o Learning in Intelligent Information Retrieval o Learning Reaction Strategies o Learning Relations o Machine Learning in Engineering Automation In addition there will be plenary talks by noted researchers in machine learning and related fields. Registration The registration fee will be $70 for students, $100 for everyone else. This fee will cover conference participation, proceedings, and receptions. To register please send a check for the appropriate amount, made out to Northwestern University, along with a completed copy of the form to be found at the end of this message. The deadline for registration is May 22, 1991. After this date, a late fee of $25 will be charged. Accomodations We have reserved rooms at the following hotels: Omni Orrington Hotel 1710 Orrington Avenue Evanston, Illinois 60201 (708) 866-8700 or 1-800-THE-OMNI Rates: $78.00 single per night $88.00 double per night The Orrington is a five minute walk from the conference site. Holiday Inn -- Evanston 1501 Sherman Avenue Evanston, Illinois 60201 (708) 491-6400 or 1-800-HOLIDAY Rates: $60.00 single per night $70.00 double per night The Holiday Inn is a fifteen minute walk from the conference site. You must make hotel reservations yourself. When you reserve your room, please inform the hotel that you are registering for the Machine Learning Workshop. We are currently trying to arrange dorm space on campus. A message about this will be forthcoming shortly. Travel By air: Chicago has two airports: O'Hare and Midway. Most flights go to O'Hare, which is also the most convenient to Evanston. To get from O'Hare to Evanston, the following options are available: Taxi: The fare from O'Hare to Evanston should be about $25. Bus Service: Continental Air Tranport (312-454-7799) and Larry Webb Bus Service (312-866-7163) leave O'Hare for Evanston every hour from the American Airlines baggage area. Each line will take you to the Omni Orrington Hotel in Evanston. The Holiday Inn is 3 blocks south of the Orrington. Public Transportation: The CTA (1-800-972-7000) "El" trains run from O'Hare to downtown Chicago, and from Chicago to Evanston. Follow the airport signs to locate the O'Hare "El" stop. The fare is $1 per person. Take the "El" to the Washington St. stop, which is the end of the line. Transfer to a Northbound train. Take this train to the Howard St. Station. Transfer to a Northbound Evanston train. Get off at the Davis St. station in Evanston. To get to the Holiday Inn, walk east to Sherman Ave., turn right and go south two blocks. To get to the Orrington, walk east to Orrington Ave., turn left and go north one block. This trip will take at least an hour. It's reasonably safe but we don't recommend it because of the aggravation involved. By car: To get to Evanston, take Dempster St. east from either the Tri-State Tollway (I-294) or the Edens Expressway (I-94). Follow Dempster east to Chicago Ave. (one block past the "El" train overpass). Turn left on Chicago and go north. To get to the Holiday Inn, turn left on Grove St., go one block west. To get to the Orrington, turn left on Davis St., go one block west to Orrington, turn right and go one and a half blocks north to the hotel. By train: Amtrack trains stop at Union Station in downtown Chicago. Walk 4 blocks north on Canal St. to the Chicago & Northwestern Commuter Train Station. Take a Northbound train to the Davis St. station in Evanston. Directions from there to the hotels are as from the "El" above. **************************Registration Form************************** Please send this form, along with a check in the appropriate amount made out to Northwestern University, to the following address: Machine Learning 1991 The Institute for the Learning Sciences 1890 Maple Avenue Evanston, Illinois, 60201 USA phone (708) 491-3500 fax (708) 491-5258 email ml91 at ils.nwu.edu Registration information (please type or print): Name: Address: Phone: Email: Type of registration: ( ) Student -- $70 ( ) Others -- $100 Registration is due May 22, 1991. If your registration will arrive after that date, please add a late fee of $25. From collins%z.ils.nwu.edu at VM.TCS.Tulane.EDU Thu Apr 11 12:28:13 1991 From: collins%z.ils.nwu.edu at VM.TCS.Tulane.EDU (Gregg Collins) Date: Thu, 11 Apr 91 11:28:13 CDT Subject: Conference announcement: ML91 Message-ID: <9104111628.AA03045@z.ils.nwu.edu> ML91 -- The Eighth International Workshop on Machine Learning Registration Information On behalf of the organizing committee, and the individual workshop committees, we are pleased to announce that ML91, the Eighth International Workshop on Machine Learning, will be held at Northwestern University, Evanston, Illinois, USA, June 27-29, 1991. ML91 comprises eight workshop tracks: o Automated Knowledge Acquisition o Computational Models of Human Learning o Constructive Induction o Learning from Theory and Data o Learning in Intelligent Information Retrieval o Learning Reaction Strategies o Learning Relations o Machine Learning in Engineering Automation In addition there will be plenary talks by noted researchers in machine learning and related fields. Registration The registration fee will be $70 for students, $100 for everyone else. This fee will cover conference participation, proceedings, and receptions. To register please send a check for the appropriate amount, made out to Northwestern University, along with a completed copy of the form to be found at the end of this message. The deadline for registration is May 22, 1991. After this date, a late fee of $25 will be charged. Accomodations We have reserved rooms at the following hotels: Omni Orrington Hotel 1710 Orrington Avenue Evanston, Illinois 60201 (708) 866-8700 or 1-800-THE-OMNI Rates: $78.00 single per night $88.00 double per night The Orrington is a five minute walk from the conference site. Holiday Inn -- Evanston 1501 Sherman Avenue Evanston, Illinois 60201 (708) 491-6400 or 1-800-HOLIDAY Rates: $60.00 single per night $70.00 double per night The Holiday Inn is a fifteen minute walk from the conference site. You must make hotel reservations yourself. When you reserve your room, please inform the hotel that you are registering for the Machine Learning Workshop. We are currently trying to arrange dorm space on campus. A message about this will be forthcoming shortly. Travel By air: Chicago has two airports: O'Hare and Midway. Most flights go to O'Hare, which is also the most convenient to Evanston. To get from O'Hare to Evanston, the following options are available: Taxi: The fare from O'Hare to Evanston should be about $25. Bus Service: Continental Air Tranport (312-454-7799) and Larry Webb Bus Service (312-866-7163) leave O'Hare for Evanston every hour from the American Airlines baggage area. Each line will take you to the Omni Orrington Hotel in Evanston. The Holiday Inn is 3 blocks south of the Orrington. Public Transportation: The CTA (1-800-972-7000) "El" trains run from O'Hare to downtown Chicago, and from Chicago to Evanston. Follow the airport signs to locate the O'Hare "El" stop. The fare is $1 per person. Take the "El" to the Washington St. stop, which is the end of the line. Transfer to a Northbound train. Take this train to the Howard St. Station. Transfer to a Northbound Evanston train. Get off at the Davis St. station in Evanston. To get to the Holiday Inn, walk east to Sherman Ave., turn right and go south two blocks. To get to the Orrington, walk east to Orrington Ave., turn left and go north one block. This trip will take at least an hour. It's reasonably safe but we don't recommend it because of the aggravation involved. By car: To get to Evanston, take Dempster St. east from either the Tri-State Tollway (I-294) or the Edens Expressway (I-94). Follow Dempster east to Chicago Ave. (one block past the "El" train overpass). Turn left on Chicago and go north. To get to the Holiday Inn, turn left on Grove St., go one block west. To get to the Orrington, turn left on Davis St., go one block west to Orrington, turn right and go one and a half blocks north to the hotel. By train: Amtrack trains stop at Union Station in downtown Chicago. Walk 4 blocks north on Canal St. to the Chicago & Northwestern Commuter Train Station. Take a Northbound train to the Davis St. station in Evanston. Directions from there to the hotels are as from the "El" above. **************************Registration Form************************** Please send this form, along with a check in the appropriate amount made out to Northwestern University, to the following address: Machine Learning 1991 The Institute for the Learning Sciences 1890 Maple Avenue Evanston, Illinois, 60201 USA phone (708) 491-3500 fax (708) 491-5258 email ml91 at ils.nwu.edu Registration information (please type or print): Name: Address: Phone: Email: Type of registration: ( ) Student -- $70 ( ) Others -- $100 Registration is due May 22, 1991. If your registration will arrive after that date, please add a late fee of $25. From cuny91 at prodigal.psych.rochester.edu Thu Apr 11 15:27:43 1991 From: cuny91 at prodigal.psych.rochester.edu (A. Zukowski (CUNY Coordinator)) Date: Thu, 11 Apr 91 15:27:43 EDT Subject: No subject Message-ID: <9104111927.AA19652@prodigal.psych.rochester.edu> ANNOUNCEMENT Language Sciences Conference Weekend at the University of Rochester May 9 - 12, 1991 There will be three overlapping conferences: The 1991 CUNY Sentence Processing Conference May 9-11 Workshop on Japanese Linguistics May 10-12 "Belief and Belief Attribution" Philosophy Conference May 11-12 ==> Conference fares are available from USAIR. <== ==> Travel scholarships are available for students. <== We will arrange group vans from nearby cities. **** If you are interested in more information about schedule **** or registration details, please forward a simple YES to: cuny91 at prodigal.psych.rochester.edu **** If you plan to attend ANY events please let us know. **** This is the second of three messages sent to newsgroups; a final conference schedule will be posted next week. From noordewi at cs.rutgers.edu Fri Apr 12 13:00:24 1991 From: noordewi at cs.rutgers.edu (noordewi@cs.rutgers.edu) Date: Fri, 12 Apr 91 13:00:24 EDT Subject: Neural Network Seminar Message-ID: <9104121700.AA08531@porthos.rutgers.edu> RUTGERS UNIVERSITY Dept. of Computer Science/Dept. of Mathematics Neural Networks Colloquium Series --- Spring 1991 C. L. Giles NEC Research Institute Teaching Recurrent Neural Networks to be Finite State Machines (Digraphs) Abstract Recurrent neural networks are natural models for encoding and learning temporal sequences. If these temporal sequences are strings from the languages of formal grammars, then teaching a neural network to learn these sequences is a form of grammatical inference. We demonstrate how to train second-order recurrent networks with real-time learning algorithms to be finite state machines. In particular, we present extensive simulation results which show that simple regular grammars are easy to learn. We devise and use heuristic clustering algorithms which extract finite state machines or digraphs from recurrent neural networks during and after training. The resultant finite state machines usually have large numbers of states and can be reduced in complexity to minimal finite state machines using a standard minimization algorithm. Depending on the training method and type of training set, different minimal finite state machines emerge. If the grammar is well learned, then identical finite state machines are produced in the minimization process. These finite state machines constitute an equivalence class of neural networks which covers different numbers of neurons and different initial conditions. This can be interpreted as a measure of how well a set of strings and its generative grammar are learned. We present a video of the learning process and show the emergent finite state machines during and after training. April 17, 1991 Busch Campus --- 4:30 p.m., room 217 SEC host: Mick Noordewier (201/932-3698) finger noordewi at cs.rutgers.edu for further schedule information From gasser at bend.UCSD.EDU Sat Apr 13 02:13:10 1991 From: gasser at bend.UCSD.EDU (Michael Gasser) Date: Fri, 12 Apr 91 23:13:10 PDT Subject: TR on development of perceptual comparison Message-ID: <9104130613.AA02880@bend.UCSD.EDU> ********************************************************************* DO NOT FORWARD TO OTHER LISTS ********************************************************************* The following paper has been placed in the neuroprose archives at Ohio State University: Comparison, Categorization, and Perceptual Dimensions: A Connectionist Model of the Development of the Notion of Sameness Michael Gasser Linda B. Smith Computer Science Department Psychology Department Indiana University Indiana University Cognitive Science Research Report 41 Abstract The ability to compare two objects is fundamental to intelligent behavior. However, it is important to distinguish between the implicit comparison that plays a role in all categorization and explicit comparison, by which two object representations are compared in short-term memory. Children learn early on both to categorize and to perform explicit comparison, but it requires a long time for them to learn to categorize and compare along particular dimensions. In this paper we present a connectionist model which brings together categorization and comparison, focusing on the development of the use of dimensions in both processes. The model posits a very general comparison mechanism which is blind to the nature of its inputs and sharing of internal object and dimension representations by categorization and comparison processes. As the system is trained on the two processes, it learns to use dimension inputs as filters on the internal representations for input objects; it is these filtered representations which are matched in comparison. The model provides a natural account of the tendency for early comparison along a single dimension to be disrupted by similarities along other, irrelevant dimensions and of the process through which the child might overcome this deficiency. ************************ How to obtain a copy ************************ a) via FTP: % ftp cheops.cis.ohio-state.edu ... Name (cheops.cis.ohio-state.edu:your-id): anonymous ... Password: neuron ... ftp> cd pub/neuroprose ... ftp> binary ... ftp> get gasser.same.ps.Z ... ftp> quit % uncompress gasser.same.ps.Z % lpr -P gasser.same.ps b) via postal mail (only if the above is impossible, please) Request a hardcopy from Beverly Brown, bebbrown at ucs.indiana.edu Cognitive Science Program, Psychology 376 Indiana University Bloomington, IN 47405 USA From tsejnowski at UCSD.EDU Sun Apr 14 01:57:22 1991 From: tsejnowski at UCSD.EDU (Terry Sejnowski) Date: Sat, 13 Apr 91 22:57:22 PDT Subject: Neural Compuation Vol 3 Issue 1 Message-ID: <9104140557.AA23939@sdbio2.UCSD.EDU> NEURAL COMPUTATION - Volume 3 Issue 1 - Spring 1991 Review: Deciphering the Brain's Codes Masakazu Konishi Letters: Synchronization of Bursting Action Potential Discharge in a Model Network of Neocortical Neurons Paul Bush and Rodney Douglas Parallel Activation of Memories in an Oscillatory Neural Network D. Horn and M. Usher Organization of Binocular Pathways: Modeling and Data Related to Rivalry Sidney R. Lehky Dynamics and Formation of Self-Organizing Maps Jun Zhang A Method for Reducing Computation in Networks with Separable Radial Basis Functions Terrence D. Sanger Adaptive Mixtures of Local Experts Robert A. Jacobs, Michael I. Jordan, Steven J. Nowlan, and Geoffrey E. Hinton Efficient Training of Artificial Neural Networks for Autonomous Navigation Dean A. Pomerleau Sequence Manipulation Using Parallel Mapping David S. Touretzky and Deirdre W. Wheeler Parsing Complex Sentences with Structured Connectionist Networks Ajay N. Jain Rules and Variables in Neural Nets Venkat Ajjanagadde and Lokendra Shastri TAG: A Neural Network Model for Large-Scale Optical Implementation Hyuek-Jae Lee, Soo-Young Lee, and Sang-Yung Shin SUBSCRIPTIONS - VOLUME 3 ______ $35 Student ______ $55 Individual ______ $110 Institution Add $18. for postage and handling outside USA (Back issues are available for $28 each.) MIT Press Journals, 55 Hayward Street, Cambridge, MA 02142. (617) 253-2889. MasterCard and VISA accepted ----- From jm2z+ at andrew.cmu.edu Tue Apr 16 09:34:59 1991 From: jm2z+ at andrew.cmu.edu (Javier Movellan) Date: Tue, 16 Apr 91 09:34:59 -0400 (EDT) Subject: Chernoff faces Message-ID: <4c2jq3C00UzxQ1J0VJ@andrew.cmu.edu> Does anybody know of stats packages with a Chernoff faces facility ? This is a method to represent multivariate observations as human faces where each variable controls a particular feature of the face. Thanks Javier Movellan From brp at bandit.berkeley.edu Tue Apr 16 13:52:03 1991 From: brp at bandit.berkeley.edu (Bruce Raoul Parnas) Date: Tue, 16 Apr 91 10:52:03 PDT Subject: models for real neural spike initiators Message-ID: <9104161752.AA00282@bandit.berkeley.edu> I am currently exploring various models used in computer simulations of real neurons, particularly implementations of spike initiators. i have been looking at the two time- constant model proposed by Hill and others many years ago, a temperature-scaled version of the Hodgkin-Huxley model, and the model proposed by Frankenhauser and Dodge. Are people aware of other spike initiator models being used in computer models of biological neurons? thanx, bruce (brp at bandit.berkeley.edu) From wray at ptolemy.arc.nasa.gov Tue Apr 16 21:06:06 1991 From: wray at ptolemy.arc.nasa.gov (Wray Buntine) Date: Tue, 16 Apr 91 18:06:06 PDT Subject: Going to IJCAI-91? Here's a relevant workshop. Message-ID: <9104170106.AA23498@ptolemy.arc.nasa.gov> Some of the nice work on engineering network structures for a given problem, multiple networks, etc., etc., would seem suitable for the following workshop. ----------------------- Call for Papers: IJCAI-91 Workshop Evaluating and Changing Representation in Machine Learning The influence of knowledge representation for the success of machine learning systems is well known. The representation language for a learning application is formed by the vocabulary (signature) as well as by the restrictions on the chosen formalism (e.g.\ number of literals). In practical applications the design of the representation for input, available domain knowledge, expert interaction and output can be time-consuming and critical for success. This workshop will focus on studies that empirically or theoretically evaluate representations, their change, and the construction of new terms. Areas of interest include: logical, relational, probabilistic and neural network representations; concept or classifier learning, model-building, knowledge-intensive learning, theory refinement, and other learning tasks; Bayesian, MDL, PAC or complexity theory; and methodologies for acquisition, evaluation and presentation of different representations and their individual components. The workshop will be held on Saturday, 24 August. Submissions (extended abstracts or short papers, 8-15 pages, 1 copy) should be sent to either of the Program Chairpersons. The collection of accepted papers will be provided to participants, and later collective publication of selected papers will be organised. CHAIRPERSONS Katharina Morik Francesco Bergadano GMD University of Torino GMD - PO Box 1240 Dipartimento di Informatica Schloss Birlinghoven corso Svizzera 185 D-5205 Sankt Augustin 1, Germany Torino, 10149, Italy (+49) 2241 14 2670, fax. 14 2889 (+39) 11 7712002, fax. 11 751603 morik at gmdzi.gmd.de bergadan at di.unito.it Wray Buntine RIACS and AI Research Branch NASA Ames Research Center Mail Stop 244-17 Moffett Field, CA, 94035, USA (+1) 415 6043389, fax. 6046997 wray at ptolemy.arc.nasa.gov IMPORTANT DATES Submission: May 15th Acceptance notified: June 15th Final camera ready copy of full paper: July 15th Workshop: August 24th (Sat.) From gaudiano at park.bu.edu Wed Apr 17 09:52:18 1991 From: gaudiano at park.bu.edu (gaudiano@park.bu.edu) Date: Wed, 17 Apr 91 09:52:18 -0400 Subject: models for real neural spike initiators In-Reply-To: connectionists@c.cs.cmu.edu's message of 16 Apr 91 20:25:12 GM Message-ID: <9104171352.AA06552@copley.bu.edu> >>>>> On 16 Apr 91 20:25:12 GMT, connectionists at c.cs.cmu.edu said: brp> From: Bruce Raoul Parnas brp> I am currently exploring various models used in computer simulations of real neurons, brp> particularly implementations of spike initiators. i have been looking at the two time- brp> constant model proposed by Hill and others many years ago, a temperature-scaled version brp> of the Hodgkin-Huxley model, and the model proposed by Frankenhauser and Dodge. brp> Are people aware of other spike initiator models being used in computer models of brp> biological neurons? A good reference book that covers this topic is: MacGregor, R.J. (1987) "Neural and Brain Modeling". Academic Press. You will find that Chapter 2 is dedicated mostly to the kind of modeling you are interested in, and includes an extensive bibliography. Paolo Gaudiano Department of Cognitive and Neural Systems Boston University 111 Cummington Street Boston, MA 02215 Phone: (617) 353-6181 e-mail: gaudiano at park.bu.edu (ARPANET) From peterc at chaos.cs.brandeis.edu Wed Apr 17 20:23:38 1991 From: peterc at chaos.cs.brandeis.edu (Peter Cariani) Date: Wed, 17 Apr 91 20:23:38 edt Subject: Spike initiation & time-dependent threshold changes In-Reply-To: gaudiano@park.bu.edu's message of Wed, 17 Apr 91 09:52:18 -0400 <9104171352.AA06552@copley.bu.edu> Message-ID: <9104180023.AA28043@chaos.cs.brandeis.edu> Regarding models for spike initiation, as far as I can see there are few (if any) models which take into account the intrinsic time course of threshold following an action potential. Thresholds for many different kinds of axons have a superexcitability period in which the threshold (that amount of injected charge needed to re-excite the membrane) descends to a half or third or less of its resting value. (See Chung, Raymond, & Lettvin, 1970; Raymond & Lettvin, 1978; Raymond, 1979). Furthermore, these thresholds are altered by activity, so the phenomenon of conduction block would serve to dynamically modify the effective connectivity between neurons (in the frequency/interspike interval domain). I don't understand why this phenomenon has been so overlooked by the neuroscience community at large, but I would think it would be extremely important for any models where "intrinsic oscillations" play an important role. I am currently working on an extremely rudimentary model of this sort, trying to understand its functional implications. It's unclear what causes this superexcitable phase--one set of ideas involves electrostatic interaction between calcium and ion channels on the outside of membranes (and is therefore related to Hill's model for accomodation)--but I think it means that neurons have more complex temporal characteristics than has been generally assumed and that oscillatory behavior may not necessarily require recurrent loops with time delays. Obviously this affects the plausibility of temporally-based pulse codes, since what required a reverberatory loop may now only need a single neuron. This all remains to be seen, but it's worth keeping in mind. SH Chung, SA Raymond, & JY Lettvin (1970) Multiple meaning in single visual neurons. Brain Behav Evol 3: 72-101 SA Raymond & JY Lettvin (1978) Aftereffects of activity in peripheral axons as a clue to nervous coding. Physiology and Pathophysiology of Axons S Waxman, ed. Raven Press, NY SA Raymond (1979) Effects of nerve impulses on threshold of frog sciatic nerve fibres. J Physiol 290:273-303. --Peter Cariani From Alexis_Manaster_Ramer at MTS.cc.Wayne.edu Wed Apr 17 22:47:46 1991 From: Alexis_Manaster_Ramer at MTS.cc.Wayne.edu (Alexis_Manaster_Ramer@MTS.cc.Wayne.edu) Date: Wed, 17 Apr 91 22:47:46 EDT Subject: MOL2 notice Message-ID: <321085@MTS.cc.Wayne.edu> ---(Forwarded from: wsavitch at UCSD.EDU, Dated: Wed, 17 Apr 91 20:16:24 EDT)--- From wsavitch at UCSD.EDU Wed Apr 17 20:16:24 1991 From: wsavitch at UCSD.EDU (wsavitch@UCSD.EDU) Date: Wed, 17 Apr 91 17:16:24 PDT Subject: No subject Message-ID: <9104180016.AA11414@walts.ucsd.edu> The Workshop on Mathematics of Language will take place on May 17 and 18 in The Tarrytown Hilton Inn in Tarrytown, New York (just outside of Yorktown Heights, N.Y). The workshop will begin at 9:30 am on May 17 and end at approximately 6:30 pm on May 18. Please note that attendance will be restricted to 80 parti- cipants. Registration will be on a first come, first served basis. To preregister please send a note to MOL2 at WATSON.IBM.COM with the following information: :name. :institution. :e-mail. :phone. If we have not run out of space by then, you will be asked to send in a check for $40 (registration at the door will be $50). All payments will have to be by check. The address of the hotel is: The Tarrytown Hilton Inn 455 South Broadway Tarrytown, New York 10591 The price of a room is $99/night. To reserve a room please call: 914-631-5700 or 1-800-HILTONS and refer to: Association for Computational Linguistics Special Interest Group on the Mathematics of Language The hotel will provide instructions on how to get there. If you have any questions, please contact: Lisa Braden-Harder (MOL 2 -- Local Arrangements) e-mail: MOL2 at WATSON.IBM.COM tel. (914) 784-7849 PROGRAM SCHEDULE FRIDAY, MAY 17, 1991 9:30 A. Manaster-Ramer Opening Address (Capacity, Complexity, and Beyond) 10:00 A. Kornai The Generative Power of Feature Geometry 10:30 BREAK 10:50 J. O'Neil The Generation of the Elementary Tree Set for a Tree-Adjoining Grammar 11:20 Mona Singh The Perfective Aspect: An Algebraic Analysis 11:50 N. Correa Attribute and Unification Grammar: A Review and Analysis of Formalisms 12:20 LUNCH 1:45 D. Radzinski Chinese Distributive Numerals and Indexed Grammars 2:15 Munidar Singh Towards a Formal Semantics for Speech Acts 2:45 BREAK 3:15 M. A. Moshier On Completeness Theorems for Feature Logics 3:45 E. Ristad On the Complexity of Two Problems in Human Language 4:15 BREAK 4:35 A. Joshi and Y. Schabes Fixed and Flexible Phrase Structure: Coordination in Tree Adjoining Grammars 5:05 DINNER 7:30 M. Dymetman A Generalized Greibach Normal Form for Definite Clause Grammars and the Decidability of the Offline-Parsability Problem 8:00 J. Nerbonne A Feature-Based Syntax/Semantics Interface 8:30 D. Johnson & L. Moss Languages Generated by Multistratal Axiomatic Grammars SATURDAY, MAY 18, 1991 9:30 R. Berwick Asymptotic Computational Complexity and Average Case Computational Complexity for Natural Language Analysis 10:00 R. Zuber An Algebraic Approach to Negation 10:30 BREAK 10.50 S. Zeitman Somewhat Finite Approaches to Infinite Sentences From tgd at turing.CS.ORST.EDU Thu Apr 18 00:14:11 1991 From: tgd at turing.CS.ORST.EDU (Tom Dietterich) Date: Wed, 17 Apr 91 21:14:11 PDT Subject: Preprint: Error-correcting output codes Message-ID: <9104180414.AA12756@turing.CS.ORST.EDU> ** Please do not forward to other newsgroups ** The following preprint is available for FTP access from the Ohio State Neuroprose archive: Error-Correcting Output Codes: A General Method for Improving Multiclass Inductive Learning Programs Thomas G. Dietterich and Ghulum Bakiri Department of Computer Science Oregon State University Corvallis, OR 97331-3202 Multiclass learning problems involve finding a definition for an unknown function f(x) whose range is a discrete set containing k>2 values (i.e., k ``classes''). The definition is acquired by studying large collections of training examples of the form . Existing approaches to this problem include (a) direct application of multiclass algorithms such as the decision-tree algorithms ID3 and CART, (b) application of binary concept learning algorithms to learn individual binary functions for each of the k classes, and (c) application of binary concept learning algorithms with distributed output codes such as those employed by Sejnowski and Rosenberg in the NETtalk system. This paper compares these three approaches to a new technique in which BCH error-correcting codes are employed as a distributed output representation. We show that these error-correcting output representations improve the performance of ID3 on the NETtalk task and of backpropagation on an isolated-letter speech-recognition task. These results demonstrate that error-correcting output codes provide a general-purpose method for improving the performance of inductive learning programs on multiclass problems. This paper will appear in AAAI-91. To retrieve: unix> ftp cheops.cis.ohio-state.edu (or 128.146.8.62) Name: anonymous Password: your-name at your-address ftp> cd pub/neuroprose ftp> binary ftp> get dietterich.error-correcting.ps.Z ftp> bye unix> uncompress dietterich.error-correcting.ps.Z unix> lpr -P{postscript printer} dietterich.error-correcting.ps Thank-you to Jordan Pollack for providing the ftp services. --Tom From brp at bandit.berkeley.edu Thu Apr 18 12:10:55 1991 From: brp at bandit.berkeley.edu (Bruce Raoul Parnas) Date: Thu, 18 Apr 91 09:10:55 PDT Subject: references to spike initiators Message-ID: <9104181610.AA02371@bandit.berkeley.edu> there have been several requests for references to the spike initiator models i mentioned in a previous message, so, here they are: One of the oldest models, and still used by many (including Macgregor) as the basis for a great many models, is the system derived by A.V. Hill, the two-time-constant model: A.V. Hill (1936), "Excitation and accomodation in nerve", Proc. Royal Soc., Series B, 119: 305-353. The model of Frankenhaeuser and Dodge (actually it'd Dodge's model, modified by Frankenhaeuser and Huxley) was developed for the frog node of Ranvier: Frankenhaeuser, B. and Huxley, A.F. (1964) "The action potential in the myelinated nerve fiber of Xenopus Laevis as computed on the basis of voltage clamp data", J. Physiol. 171: 302-315. (note: the spelling may be either Frankenhaeuser or frankenhauser, depending on where you look, the ae is used to indicate an a-umlaut in the German spelling). And, of course, there is the HH model, presented in a series of papers: Hodgkin, A.E. and Huxley, A.F. in J. Physiol., 1951 and 1952. bruce (brp at bandit.berkeley.edu) From chan%unb.ca at UNBMVS1.csd.unb.ca Thu Apr 18 16:08:27 1991 From: chan%unb.ca at UNBMVS1.csd.unb.ca (Tony Chan) Date: Thu, 18 Apr 91 17:08:27 ADT Subject: Call for participation in AI Symposium Message-ID: *************************** * * * CALL for PARTICIPATION * * * *************************** The Fourth UNB Artificial Intelligence Symposium University of New Brunswick Fredericton, New Brunswick, Canada Friday & Saturday, Sept. 20 & 21, 1991 Program Committee: ****************** Ranan Banerji, St. Joseph's U., Philadelphia, PA Wolfgang Bibel, I.T.H., Darmstadt, West Germany David Bonham, Mechanical Eng., U.N.B. Z. Chen, U. of Nebraska, Omaha, NE Chang Choo, W.P.I., Worcester, MA Rajamani Doraiswami, Electrical Eng., U.N.B. Martin A. Fischler, SRI Int., Menlo Park, CA Herbert Freeman, Rutgers U., New Brunswick , NJ C. Lee Giles, NEC Research Ins., Princeton, NJ Paul Gillard, Memorial U., St. John's, Nfld. Vasant Honavar, Iowa State U., Ames, IA Laveen Kanal, U. of Maryland, Coll. Park, MD Vladik Kreinovich, U. of Texas, El Paso, TX Werner Kuhn, U. of Maine, Orono, ME Bernd Kurz, Computer Science, U.N.B. Patrice Lapointe, AECL Res., Chalk River, Ont. Mark Lidd, Mitre Corp., Fairfax, VA B.I.B. Madhav, I.I.T., Madras, India Ettore Merlo, C.R.I.M., Montreal, Que. Tom Mitchell, Carnegie Mellon, Pittsburgh, PA Jan Mulder, Dalhousie U., Halifax, N.S. Eric Neufeld, U. of Saskatchewan, Saskatoon, Sask. N. Parameswaran, I.I.T., Madras, India Richard Peacocke, Bell-Northern Res., Ottawa, Ont. Denis Riordan, T.U.N.S., Halifax, N.S. John Robinson, Civil Eng., U.N.B. Azriel Rosenfeld, U. of Maryland, Coll. Park, MD Leemseop Shim, Governors State U., Univ. Park, IL Eduardo Sontag, Rutgers U., New Brunswick, NJ Paul Tarau, U. de Moncton, Moncton, N.B. Manoel Tenorio, Purdue U., West Lafayette, IN Richard Tervo, Electrical Eng., U.N.B. Tetsuyuki Toyofuku, National-Panasonic, Osaka, Japan Andri Trudel, Acadia U., Wolfville, N.S. Lloyd Waugh, Civil Eng., U.N.B. Ian Witten , U. of Calgary, Calgary, Alta. Andrew K.C. Wong, U. of Waterloo, Waterloo, Ont. Jean Zanazaka, U. de Sherbrooke, Que. Sponsored by: ************* Faculty of Computer Science, University of New Brunswick Bell-Northern Research Canadian Society for Computational Studies of Intelligence In Cooperation with: ******************** The American Association for Artificial Intelligence (AAAI) IEEE Computer Society International Association of Knowledge Engineers (IAKE) Japanese Society of Artificial Intelligence Keynote Address: **************** "Self-Reliant Robots: The Ambler Rover and Beyond" by Dr. Reid Simmons, Research Computer Scientist, Robotics Institute, Carnegie Mellon University, Pittsburgh, PA Panel Discussion (1.5 hours): ***************************** "On the Role of Machine Learning in Artificial Intelligence" ---------------------------------------------------------------------- The UNB Artificial Intelligence Symposium provides a forum for an exchange of ideas, experiences and information about activities in artificial intelligence. Your participation is invited in the form of (1) a paper to be reviewed (nominally by three reviewers) and, if accepted, published and presented at the Symposium, or (2) a half-day tutorial on a suitable AI topic, to be presented at the Symposium. All tutorial proposals must have a stated goal, a detailed outline (2 - 5 pages), target audience defined, and a brief resume of the tutorial presenter(s), or (3) an exhibit or display of artificial intelligence technology. Important Dates: **************** May 15, 1991 -- Four copies of an extended abstract (2 to 4 double- spaced pages) or a full paper (max. 11 pages, including figures) to one of the Program Co-Chairs (see address below). July 1, 1991 -- Notification of acceptance will be mailed. August 15, 1991 -- Final camera-ready copies of papers are due. Topics of interest include, but are not limited to *************************************************** Foundations of AI, Connectionist Theory & Applications Pattern Recognition, Machine Learning Knowledge Acquisition, Automated Theorem Proving Computer Vision, Knowledge Representation Natural Language Processing , Temporal Reasoning Knowledge-Based Systems, Database/Knowledge Base Integration Robotics, Real-Time Expert Systems, Automated Planning Expert System Verification, Automated Scheduling Novel Architectures for AI, Case-Based Reasoning Truth Maintenance Systems, Reasoning Under Uncertainty A pre-published proceedings will be available at the Symposium. Submit papers or extended abstracts to: Program Co-Chairs, 4th UNB AI Symposium Brad Nickerson or Lev Goldfarb E-mail: bgn at unb.ca or goldfarb at unb.ca Submit tutorial proposals to: Bruce Spencer or Przemyslaw Pochec Tutorial Co-Chairs, 4th UNB AI Symposium E-mail: bspencer at unb.ca or pochec at unb.ca Submit exhibit requests to: Kirby Ward Exhibits Chair, 4th UNB AI Symposium E-mail: wardk at unb.ca The mailing address, phone and fax numbers for the above are as follows: Faculty of Computer Science, University of New Brunswick P.O. Box 4400, Fredericton, N.B., Canada E3B 5A3 Phone: (506) 453-4566 Fax: (506) 453-3566 It is planned to hold the CKEd (Certified Knowledge Engineer) examination on Sept. 19, 1991 at the Symposium location. Symposium participants wishing to take the examination should contact the IAKE at International Associaton of Knowledge Engineers Georgetown P.O. Box 25461, Washington, D.C. 20007, U.S.A. ph: 301 231 7826 fax: 301 770 4621 E-mail: IAKE at UC780.bitnet From koch at CitIago.Bitnet Thu Apr 18 16:59:13 1991 From: koch at CitIago.Bitnet (Christof Koch) Date: Thu, 18 Apr 91 13:59:13 PDT Subject: Spike initiation & time-dependent threshold changes In-Reply-To: Your message <9104180023.AA28043@chaos.cs.brandeis.edu> dated 17-Apr-1991 Message-ID: <910418135913.2040bf1d@Iago.Caltech.Edu> The fact that the threshold of firing goes up following action potential dischar ge is a well know and well studied model. It's paartially caused by the influx of calcium, which activates one or two calcium-dependent potassium currents. See the model in Chapter 4 ("Multiple channels and calcium dynamics" in the book "Methods in Neuronal Modeling", (C. Koch and I. Segev, eds.), MIT Press, 1989. This book contains a number of quite detailed, state-of-the-art models. One of the best studied models of spike-initiation is still the integrate-and-fire or leaky-integrate-and-fire models, develoepd by Brice Knight. It's quite well described in Tuckwell's two volume monograph "Introduction to theoretical neurobiology" (Cambridge Univ. Press, 1988). Some analytical solutions are known, for instance in the case that the input is a Poisson input. Christof From brp at bandit.berkeley.edu Fri Apr 19 00:44:19 1991 From: brp at bandit.berkeley.edu (Bruce Raoul Parnas) Date: Thu, 18 Apr 91 21:44:19 PDT Subject: Spike initiation & time-dependent threshold changes Message-ID: <9104190444.AA03832@bandit.berkeley.edu> One of the best studied models of spike-initiation is still the integrate-and-fire or leaky-integrate-and-fire models, develoepd by Brice ^^^^^ Knight. It's quite well described in Tuckwell's two volume monograph ^^^^^^ "Introduction to theoretical neurobiology" (Cambridge Univ. Press, 1988). Christof While i agree that simple integrate and fire models (one time constant or two) are definitely among the best around, I find it odd thatt so many people attribute these models to people other that those who developed them. The earliest work I know of on two time constant models was done in 1936 by A.V. Hill, when much less was known about neurophysiology. Rashevsky and Monier worked on similar models at about the same time, but the concept is 55 years old. bruce From cuny91 at prodigal.psych.rochester.edu Fri Apr 19 08:36:05 1991 From: cuny91 at prodigal.psych.rochester.edu (A. Zukowski (CUNY Coordinator)) Date: Fri, 19 Apr 91 08:36:05 EDT Subject: No subject Message-ID: <9104191236.AA11419@prodigal.psych.rochester.edu> PLEASE POST HARD COPY WHERE PERTINENT LANGUAGE SCIENCES WEEKEND AT THE UNIVERSITY OF ROCHESTER THERE WILL BE THREE CONFERENCES ON LANGUAGE MAY 9-12 AT THE UNIVERSITY OF ROCHESTER. ANNUAL CUNY SENTENCE PROCESSING CONFERENCE ANNUAL JAPANESE GRAMMAR CONFERENCE ROCHESTER PHILOSOPHY CONFERENCE ON BELIEF AND BELIEF ATTRIBUTION WE HAVE CRASH SITES, TRAVEL FELLOWSHIPS AND OTHER LOCAL ARRANGEMENTS AVAILABLE. IF YOU ARE INTERESTED IN ATTENDING, PLEASE CONTACT US: NAME: EMAIL: PHONE: SURFACE MAIL: SCHEDULES FOLLOW 1991 CUNY Sentence Processing Conference Schedule ------------------------------------------------- THURSDAY, MAY 9 12:55 -1:00 Opening Remarks 1:00 - 1:35 Ivan Sag "Taking Performance Seriously: A Plea for Constraint" 1:35 - 2:00 Weijia Ni & Stephen Crain "Avoiding Garden Paths" 2:00 - 2:25 Gerry Altmann, Alan Garnham, Yvette Dennis, & A. Henstra "Eye Movements Controlled: Evidence For and Against Context Effects" 2:25 - 2:50 Don Mitchell, Martin Corley, & Alan Garnham "Biases and Influences in Human Sentence Parsing: Evidence for the Precedence of "Pure" Syntactic Strategies" 2:50 - 3:15 Maryellen C. MacDonald "Pre- and Post-ambiguity Cues for Syntactic Ambiguity Resolution" 3:15 - 4:00 Break 4:00 - 5:30 Japanese Psycholinguistics Symposium (Reiko Mazuka, Mineharu Nakayama, Yoshi Otsu, Peter Culicover) 5:50 - 7:30 Supper 7:30 - 9:30 Connectionism and Models of Language (Paul Smolensky, Gary Dell, Jerry Feldman) ------------------------------------------------------------------------------- FRIDAY, MAY 10 8:30 - 9:00 Coffee 9:00 - 9:25 Janet Nicol, Cecile McKee, & Dana McDaniel "The Development of the Coreference Processor" 9:25 - 9:50 Howard Kurtzman, Raul Elias-Cintron, & Rocio Aramburu "Probe Recognition in Spanish: Effects of NP-trace and Agreement Morphology" 9:50 - 10:15 Karen Emmorey & Diane Lillo-Martin "Processing Spatial Anaphora: Referent Activation with Overt and Null Pronouns in American Sign Language" 10:15 - 10:40 Kay Bock & Kathleen M. Eberhard "The Meaning, Sound, and Syntax of English Number Agreement" 10:40 - 11:00 Break 11:00 - 12:30 Second Language Acquisition Symposium (Peter Jusczyk, Jacques Mehler, Elissa Newport) 12:30 - 2:00 Lunch 2:00 - 2:25 Di Bradley "Lexical and Post-Lexical Contributions to Priming" 2:25 - 2:50 (To Be Announced) 2:50 - 3:15 Michael Brent "Automatic Acquisition of Subcategorization Frames from Untagged, Free-Text Corpora" 3:15 - 3:40 Mary Potter "Word Perception in Sentence Context: Back to the Future" 3:40 - 4:00 Break 4:00 - 5:30 Conditionals Symposium (William Lycan, Phillip Johnson-Laird) 5:30 - 7:30 POSTERS Tod A. Bassham & Jill N. Beckman "The Parsing of Tough- and Too-Constructions: Implications for Models of Gap-Location" Tom Bever, Cornell Juliano, & Andrea Zukowski "Improvement of Text Readability Through Applied Psycholinguistics" Elizabeth R. Blackmer "Implications of the Timing of Rapid Multiple Repairs in Spontaneous Speech" Giuseppe Cappelli, Anna Maccari & Lucia Pfanner "A System for Semi-Automatical Treatment of Child Morphology" Soon Ae Chun "Verbal Suffixes and Control Interpretation in Korean" Mark Coulson "Processing Ambiguous Anaphora" Aaron Halpern "Priming in VP Anaphora" Gregory Hickok, Enriqueta Canseco, Edgar Zurif, & Jane Grimshaw "Modularity in Locating Wh-gaps" Dan Jurafsky "An On-Line Model of Human Sentence Interpretation" Sheila Meltzer "(How) Does Discourse Pragmatics Influence Processing of Phonologically Null Pronouns?" Gail Mauner "Syntactic Context and the Interpretation of Verb Phrase Anaphors" Neal Perlmutter "Effects of Pragmatic Cues on Syntactic Ambiguity Processing" Justin Peterson & Dorrit Billman "I'm with Her and the Butler Did It with the Knife" Susan M. Powers "'I Give Up': Children's Processing of Sentences Containing Verb Particles" Leslie Re, Harriet Taber, & Janet D. Fodor "Reconciling Methodological Issues in Empty Category Processing" Vincent J. Samar "Be is a Raising Verb: Psycholinguistic Evidence" Yoshinori Sasaki "Development of Sentence Processing Strategies in English and Japanese as Foreign Languages: An Analysis Based on the Competition Model" Jan C. Scholtes "Kohonen's Self Organizing Map Applied Towards Natural Language Processing" Julie Sedivy "The Use of Thematic Relations in Gap-Filling" L. P. Shapiro, H. N. Nagel, & B. A. Levine "Verb Frame Preferences and Sentence Processing: Implications for a Model of Parsing" Ron Smyth & Kumiko Murasugi "Children's Use of Agreement Clues to Relative Clause Attachment" Laurie A. Stowe, Peter W. Culicover, Michael Torello, Robert Angel, & Tod Bassham 7:30 - 9:30 Banquet 9:30 - Festivities ------------------------------------------------------------------------------- SATURDAY, MAY 11 9:00 - 9:30 Coffee 9:30 - 10:15 Susan Garnsey - ERP Tutorial 10:15 - 10:40 Rob Kluender & Marta Kutas "Evidence for Lexical and Syntactic Processing Effects in Wh-Islands" 10:40 - 11:00 Break 11:00 - 1:00 Reading Symposium (Marcel Just, Mark Seidenberg, Gail McKoon) 1:00 - 2:00 Lunch 2:00 - 2:25 Janet Dean Fodor "Empty Categories at S-Structure and PF" 2:25 - 2:50 Andrew Barss "The Grammar of Antecedence and Antecedent Reactivation" 2:50 - 3:15 Martin Pickering & Guy Barry "Gap-Free Processing and the Resolution of Ambiguities" 3:15 - 4:00 Edward Gibson "A Computational Treatment of Processing Overload" 4:00 - 4:05 Break 4:05 - 5:35 Lexical Concepts Symposium (Frank Keil, James Hampton) ********************** The Rochester Workshop on Japanese Linguistics, Universal Grammar, and Their Implications to Language Pedagogy and Human Cognition (tentative schedule) --------------------------------------------------------------- Friday, May 10 -------------- 9.00-9.15 Opening Remarks 9.15-10.15 On Scrambling Naoko Nemoto, University of Connecticut 10.15-10.30 Break 10.30-11.30 Scrambling and A-chains Shigeru Miyagawa, Ohio State University 11.30-12.30 Scrambling and the FOCUS Interpretation Ayumi Ueyama, Kyoto University of Foreign Studies 12.30-2.15 Lunch 2.15-3.15 Focused Pronouns Richard Larson, SUNY Stony Brook 3.15-4.15 Subject-to-Object Raising, ECM and the Major Object in Japanese Hojime Hoji, USC 4.15-4.30 Break 4.30-5.30 The Syntax of Answers: A Comparative Approach Itziar Laka, University of Rochester 7.30-9.30 Banquet (with CUNY Conference participants) Saturday, May 11 ---------------- 9.45-10.45 Reanalyzing Reanalysis Natsuko Tsujimura, University of Illinois 10.45-11.45 Title TBA Shige-Yuki Kuroda, UC San Diego 11.45-2.00 Lunch 2.00-3.00 On Passive Hiroto Hoshi, University of Connecticut 3.00-4.00 De-Passivization Yoshihisa Kitagawa, University of Rochester 4.00-4.15 Break 4.15-5.15 Two Types of Inflection David Pesetsky, MIT Sunday, May 12 -------------- 10.00-11.00 Negative Polarity Items and Floating Quantifiers Naoya Fujita, University of Rochester 11.00-12.00 Dissussion Session 1) Comments on Nemoto, Miyagawa, and Ueyama Hiroaki Tada, MIT 2) Comments on Larson Shin Watanabe,USC ***************************** The Univerisity of Rochester Department of Philosphy's Conference on Belief and Belief Attribution May 11-12, 1991 --------------------------------------------------------------------- Saturday, May 11 ---------------- 9:30 Robert Stalnaker "On Twin Earth" 11:00 Ernest LePore "Holism & Belief" --lunch-- 2:00 William Lycan "A New Argument for the Representational Theory of Thinking" 3:00 Mark Crimmins "Tacitness Doesn't Make a Difference Sunday, May 12 -------------- 9:30 James Higginbotham "Belief & Logical Form" 11:00 Scott Soames "Attitudes & Anaphora" Information about the conferences may be requested from the following sources: CUNY: Andrea Zukowski (cuny91 at prodigal.psych.rochester.edu) Psychology Department University of Rochester Rochester, NY 14627 (716) 275-0754 (716) 275-8724 Japanese Linguistics, UG Workshop: Yoshi Kitagawa (ykit%uorvm.bitnet at db1.cc.rochester.edu) Foreign Languages, Literature, & Linguistics University of Rochester Rochester, NY 14627 (716) 275-4298 (716) 275-4252 Belief and Belief Attribution Conference: David Braun (dbrn at uhura.cc.rochester.edu) Department of Philosophy Univerisity of Rochester Rochester, NY 14627 (716) 275-4105 From stolcke at ICSI.Berkeley.EDU Sat Apr 20 18:27:39 1991 From: stolcke at ICSI.Berkeley.EDU (stolcke@ICSI.Berkeley.EDU) Date: Sat, 20 Apr 91 16:27:39 MDT Subject: new neta release of cluster program Message-ID: <9104202327.AA16136@icsib30.Berkeley.EDU> Dear Connectionists: This is the second release of my enhanced version of the `cluster' program (original version by Y. Miyata). It has a bunch of added features, most importantly I merged a Principal Component Analysis program I once wrote into this program, so it now does both HCA and PCA, using compatible input and output formats. I and a couple of people around here have been using this with great success, but there are bound to be some bugs left, especially in the less commonly used features. I would encourage people who have a need for this kind of tool to give it a try and report errors and shortcomings back to me. Once I'm confident that there are no major deficiencies left I'll make an updated version available and announce it on this list. Andreas HOW TO GET CLUSTER cluster is available via anonymous ftp from icsi-ftp.berkeley.edu (128.32.201.55). To get it use FTP as follows: % ftp icsi-ftp.berkeley.edu Connected to icsic.Berkeley.EDU. 220 icsi-ftp (icsic) FTP server (Version 5.60 local) ready. Name (icsic.Berkeley.EDU:stolcke): anonymous Password (icsic.Berkeley.EDU:anonymous): 331 Guest login ok, send ident as password. 230 Guest login Ok, access restrictions apply. ftp> cd pub/ai 250 CWD command successful. ftp> binary 200 Type set to I. ftp> get cluster-2beta.tar.Z 200 PORT command successful. 150 Opening BINARY mode data connection for cluster-2beta.tar.Z (15531 bytes). 226 Transfer complete. 15531 bytes received in 0.08 seconds (1.9e+02 Kbytes/s) ftp> quit 221 Goodbye. HOW TO BUILD CLUSTER Unpack in an empty directory using % zcat cluster-2beta.tar.Z | tar xf - Read the README and especially the man page (cluster.man) for information. Then compile with % make After making the appropriate adjustments in the Makefile you can % make install On most BSD-based UNIX systems cluster should compile without problems. The Makefile gives some hint as to what might have to be changed to get it to run. And please, tell me where I goofed! From tsejnowski at UCSD.EDU Sat Apr 20 22:32:42 1991 From: tsejnowski at UCSD.EDU (Terry Sejnowski) Date: Sat, 20 Apr 91 19:32:42 PDT Subject: Neural Computation Long Papers Message-ID: <9104210232.AA26247@sdbio2.UCSD.EDU> The editorial board of Neural Computation has decided to start accepting a limited number of full length papers. These will be more thoroughly reviewed than letters and there will be only one or two per issue, so it will be more difficult to get one accepted, like the longer articles in Nature and Science. Terry ----- From gelenbe at csd36.NYU.EDU Mon Apr 22 12:26:56 1991 From: gelenbe at csd36.NYU.EDU (Erol Gelenbe) Date: Mon, 22 Apr 91 12:26:56 -0400 Subject: IEEE Transactions on Software Eng. Special Issue Message-ID: <9104221626.AA05267@csd36.NYU.EDU> A special isssue of the IEEE TSE on "Artificial Neural Network Models and Systems" is planned for the month of April 1992, for which I will as Editor. Papers concerning models and their theory, system implementations including one or more applications, and system environments which use artificial neural networks are sollicited. They will be reviewed in accordance with the normal practices of these IEEE Transactions. Paper submissions should be sent to me by August 1, 1991 to the following address : Erol Gelenbe EHEI 45 rue des Saints-Peres 75006 Paris, France erol at ehei.ehei.fr From fay at archsci.arch.su.OZ.AU Tue Apr 23 05:49:34 1991 From: fay at archsci.arch.su.OZ.AU (fay@archsci.arch.su.OZ.AU) Date: Tue, 23 Apr 91 19:49:34 +1000 Subject: Symposium on AI, Reasoning and Creativity Message-ID: <9104230949.AA09174@chomsky.arch.su.OZ.AU> Could you please let me know if you are interested in attending or submitting a paper to this symposium. Thanks Fay * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * PRELIMINARY ANNOUNCEMENT AND CALL FOR PAPERS * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * SYMPOSIUM ON ARTIFICIAL INTELLIGENCE, REASONING AND CREATIVITY 20-23 AUGUST 1991 immediately preceding the International Joint Conference on Artificial Intelligence (IJCAI'91) organised by GRIFFITH UNIVERSITY, BRISBANE, AUSTRALIA to be held at LAMINGTON NATIONAL PARK, QUEENSLAND, AUSTRALIA AIM Creativity is one of the least understood aspects of intelligence. It is commonly regarded as 'intuitive' and not susceptible to rational enquiry. However, there is now considerable work in artificial intelligence and cognitive science which addresses creativity. This symposium will provide a forum for exploring and discussing these ideas, and for suggesting directions for future research. It aims to attract practitioners of both 'cognitive' and 'technological' artificial intelligence. KEYNOTE SPEAKER - PROFESSOR MARGARET BODEN Margaret Boden, in her talk on `Creativity and Computers', will discuss how computational concepts drawn from artificial intelligence can explore creativity. Computers can sometimes do apparently creative things; more to the point, they can suggest how we manage to do so. Computational ideas are therefore helping us to understand how human originality is possible. Margaret Boden is Professor of Philosophy and Psychology, and Founding Dean of the School of Cognitive Sciences, at the University of Sussex, UK. Her recent publications include `Artificial Intelligence and Natural Man' (1987), `Artificial Intelligence in Psychology' (1989) and `The Creative Mind: Myths and Mechanisms (1991). INVITED SPEAKERS include Ernest Edmonds, Loughborough University, UK John Gero, University of Sydney, Australia Graham Priest, University of Queensland, Australia Roger Wales, University of Melbourne, Australia TOPICS FOR PRESENTATION AND DISCUSSION The symposium calls for extended abstracts of 750-1,000 words. The following are suggested areas only: - Models of creativity - Modelling creative processes - Creative reasoning, e.g theory generation in science - Analogical reasoning - Case-based reasoning - Nonmonotonic reasoning - Thought experiments Both connectionist and traditional symbolic approaches are welcome. All abstracts will be refereed. Accepted papers will be subjected to a further refereeing process for publication by Kluwer Academic Press. The symposium will be structured to provide adequate time for both presentation and discussion. SUBMISSION Three copies of extended abstracts are required by 31 May 1991. Abstracts may be submitted electronically as LaTeX or plain ASCII files via email but hard copies must also be submitted. Two hard copies of final versions of accepted papers, and an electronic version on Macintosh disk or via email are required at the time of registration on 20 August 1991. Authors of accepted papers will be expected to sign a copyright release form to allow publication of the proceedings. At least one author of each paper is expected to present the paper at the symposium. PREPRINTS AND PROCEEDINGS Accepted extended abstracts will be printed in the form of Preprints and be available for distribution at the time of registration. Full papers will be published subsequently by Kluwer Academic Press. LOCATION The symposium is being held at O'Reillys Lodge in Lamington National Park, Queensland - a rainforest 120 km south of Brisbane. Rainforests vary from the temperate beech forest of the higher altitudes to the warm subtropical rainforest of the valleys. There is an abundance of rare and spectacular plants (orchids, ferns, giant epiphytes, mosses, luminous fungi) and a diverse community of birds and animals (crimson rosellas, king parrots, bower birds, miniature kangaroos, brush turkeys) all coexisting beneath the lofty rainforest canopy. On the final day of the symposium there will be ample time for guided or independent bush walks, 4WD bus trips, barbeques and evening entertainment. FACILITIES O'Reillys is a mountain resort of modern units. The seminar facilities are highly professional, including lecture theatre, audiovisual equipment and library. TIMETABLE Extended abstracts (750-1000 words) - 3 hard copies 31 May 1991 Notification of acceptances 17 June 1991 Full formatted papers due 20 August 1991 Symposium 20-23 August 1991 COSTS in Australian dollars (US$1 = ~A$1.28; PStg 1 = ~A$2.33) Registration fee (including one copy of Preprints): Full fee $250 Authors (1 per paper) $150 Accommodation (including all meals): Bethongabel units (private bath, balcony, view) $119 pp/pn Elabana units (private bath, limited availability) $105 pp/pn Bus to Lamington National Park from Griffith University: Round trip $25 INTERNATIONAL COMMITTEE Chair Terry Dartnall, Griffith University, Australia Conference Organiser Fay Sudweeks, University of Sydney, Australia Advisory Board Margaret Boden, Sussex University, UK Andy Clarke, Sussex University, UK Max Coltheart, Macquarie University, Australia Boi Faltings, Ecole Polytechnique Federale de Lausanne, Switzerland Marilyn Ford, Griffith University, Australia John Gero, University of Sydney, Australia Rod Girle, Griffith University; Australian National University Laveen Kanal, University of Maryland, USA Graham Priest, University of Queensland, Australia Simon Ross, University College of London, UK; Kluwer Academic Press Aaron Sloman, Sussex University, UK Roger Wales, University of Melbourne, Australia Janet Wiles, University of Queensland, Australia CONFERENCE CONTACTS Correspondence and queries: Dr Terry Dartnall School of Computing and Information Technology Griffith University Nathan Qld 4111 Australia Tel: +61-7-875 5020 Fax: +61-7-875 5198 Email: terryd at gucis.sct.gu.edu.au Abstracts and papers: Ms Fay Sudweeks Department of Architectural and Design Science University of Sydney NSW 2006 Australia Tel: +61-2-692 2328 Fax: +61-2-692 3031 Email: fay at archsci.arch.su.oz.au Registration: Ms Denise Vercoe School of Computing and Information Technology Griffith University Nathan Qld 4111 Australia Tel: +61-7-875 5002 Fax: +61-7-875 5198 From collins at z.ils.nwu.edu Tue Apr 23 12:58:23 1991 From: collins at z.ils.nwu.edu (Gregg Collins) Date: Tue, 23 Apr 91 11:58:23 CDT Subject: On-campus housing for ML-91 Message-ID: <9104231658.AA00525@z.ils.nwu.edu> We are pleased to announce the availability of on-campus housing for ML91, the Eighth International Workshop on Machine Learning, which will be held at Northwestern University, Evanston, Illinois, USA, June 27-29, 1991. Housing will consist of double-occupancy dormitory rooms at a rate of $69.00 for four nights (June 26-29). *****************ML91 On-Campus Housing Registration Form***************** Please send this form, along with a check for $69.00 made out to Northwestern University, to the following address: Machine Learning 1991 The Institute for the Learning Sciences 1890 Maple Avenue Evanston, Illinois, 60201 USA phone (708) 491-3500 fax (708) 491-5258 email ml91 at ils.nwu.edu Registration information (please type or print): Name: Sex: Address: Phone: Email: Name of person with whom you would like to share your room (if left blank, we will assign you a roommate): From collins%z.ils.nwu.edu at VM.TCS.Tulane.EDU Tue Apr 23 12:58:23 1991 From: collins%z.ils.nwu.edu at VM.TCS.Tulane.EDU (Gregg Collins) Date: Tue, 23 Apr 91 11:58:23 CDT Subject: On-campus housing for ML-91 Message-ID: <9104231658.AA00525@z.ils.nwu.edu> We are pleased to announce the availability of on-campus housing for ML91, the Eighth International Workshop on Machine Learning, which will be held at Northwestern University, Evanston, Illinois, USA, June 27-29, 1991. Housing will consist of double-occupancy dormitory rooms at a rate of $69.00 for four nights (June 26-29). *****************ML91 On-Campus Housing Registration Form***************** Please send this form, along with a check for $69.00 made out to Northwestern University, to the following address: Machine Learning 1991 The Institute for the Learning Sciences 1890 Maple Avenue Evanston, Illinois, 60201 USA phone (708) 491-3500 fax (708) 491-5258 email ml91 at ils.nwu.edu Registration information (please type or print): Name: Sex: Address: Phone: Email: Name of person with whom you would like to share your room (if left blank, we will assign you a roommate): From harnad at Princeton.EDU Tue Apr 23 16:31:25 1991 From: harnad at Princeton.EDU (Stevan Harnad) Date: Tue, 23 Apr 91 16:31:25 EDT Subject: 2 TRs: Categorical Perception and Neural Nets Message-ID: <9104232031.AA00250@reason.Princeton.EDU> The following two tech reports are available by anonymous ftp from directory /pub/harnad on princeton.edu. Full ftp instructions follow the abstracts. ------------------------------------------------- (1) Categorical Perception and the Evolution of Supervised Learning in Neural Nets S Harnad*, SJ Hanson*,** & J Lubin* *Princeton University **Siemens Research Center [Presented at 1991 AAAI Symposium on Symbol Grounding: Problem and Practice] ABSTRACT: Some of the features of animal and human categorical perception (CP) for color, pitch and speech are exhibited by neural net simulations of CP with one-dimensional inputs: When a backprop net is trained to discriminate and then categorize a set of stimuli, the second task is accomplished by "warping" the similarity space (compressing within-category distances and expanding between-category distances). This natural side-effect also occurs in humans and animals. Such CP categories, consisting of named, bounded regions of similarity space, may be the ground level out of which higher-order categories are constructed; nets are one possible candidate for the mechanism that learns the sensorimotor invariants that connect arbitrary names (elementary symbols?) to the nonarbitrary shapes of objects. This paper examines how and why such compression/expansion effects occur in neural nets. [Retrieve by anonymous ftp in binary mode as (compressed) file harnad91.cpnets.Z from directory /pub/harnad on princeton.edu, instructions below] ----------------------------------------------------------------- (2) Connecting Object to Symbol in Modeling Cognition Stevan Harnad Department of Psychology Princeton University Princeton NJ 08544 [To appear in Clark, A. & Lutz, R. (Eds) (1992) "CONNECTIONISM IN CONTEXT," Springer-Verlag] Connectionism and computationalism are currently vying for hegemony in cognitive modeling. At first glance the opposition seems incoherent, because connectionism is itself computational, but the form of computationalism that has been the prime candidate for encoding the "language of thought" has been symbolic computationalism, whereas connectionism is nonsymbolic. This paper examines what is and is not a symbol system. A hybrid nonsymbolic/symbolic system will be sketched in which the meanings of the symbols are grounded bottom-up in the system's capacity to discriminate and identify the objects they refer to. Neural nets are one possible mechanism for learning the invariants in the analog sensory projection on which successful categorization is based. "Categorical perception," in which similarity space is "warped" in the service of categorization, turns out to be exhibited by both people and nets, and may mediate the constraints exerted by the analog world of objects on the formal world of symbols. [Retrieve by anonymous ftp in binary mode as (compressed) file harnad92.symbol.object.Z from directory /pub/harnad on princeton.edu] To retrieve a file by ftp from a Unix/Internet site, type: ftp princeton.edu When you are asked for your login, type: anonymous For your password, type your full name then change directories with: cd pub/harnad Then type: binary (This is for retrieving compressed files.) To show the available files, type: ls Next, retrieve the file you want with (for example): get filename.Z When you have the file(s) you want, type: quit Next uncompress the file with: uncompress filename.Z Now the file will be called, simply, filename --- The above cannot be done from Bitnet directly, but there is a fileserver called bitftp at pucc.bitnet that will do it for you. Send it the one line message help for instructions (which will be similar to the above, but will be in the form of a series of lines in an email message that bitftp will then execute for you). From sankar at bach.rutgers.edu Tue Apr 23 18:49:46 1991 From: sankar at bach.rutgers.edu (sankar@bach.rutgers.edu) Date: Tue, 23 Apr 91 18:49:46 EDT Subject: No subject Message-ID: <9104232249.AA05548@bach.rutgers.edu> The following two papers are now available via FTP from the neuroprose archives. Both will be presented at IJCNN-Seattle, 1991. These papers describe a new approach that combines Neural Networks and Decision Trees to form a classifier that grows the neurons as it learns. **************************************************************************** SPEAKER INDEPENDENT VOWEL RECOGNITION USING NEURAL TREE NETWORKS Ananth Sankar and Richard Mammone CAIP Center and Dept. of Electrical Engg. Rutgers University, P.O. Box 1390 Piscataway, NJ 08855-1390 Speaker independent vowel recognition is a difficult pattern recognition problem. Recently there has been much research using Multi-Layer Perceptrons (MLP) and Decision Trees for this task. This paper presents a new approach to this problem. A new neural architecture and learning algorithm called Neural Tree Networks (NTN) are developed. This network uses a tree structure with a neural network at each tree node. The NTN architecture offers a very efficient hardware implementation as compared to MLPs. The NTN algorithm grows the neurons while learning as opposed to backpropagation, for which the number of neurons must be known before learning can begin. The new algorithm is guaranteed to converge on the training set whereas backpropagation can get stuck in local minima. A gradient descent technique is used to grow the NTN. This approach is more efficient than the exhaustive search techniques used in standard decision tree algorithms. We present simulation results on a speaker independent vowel recognition task. These results show that the new method is superior to both MLP and decision tree methods. ***************************************************************************** OPTIMAL PRUNING OF NEURAL TREE NETWORKS FOR IMPROVED GENERALIZATION Ananth Sankar and Richard Mammone CAIP Center and Dept. of Electrical Engg. Rutgers University, P.O. Box 1390 Piscataway, NJ 08855-1390 An optimal pruning algorithm for a Neural Network recently developed called Neural Tree Networks (NTN) is presented. The NTN is grown by a constructive learning algorithm that decreases the classification error on the training data recursively. The optimal pruning algorithm is then used to improve generalization. The pruning algorithm is shown to be computationally inexpensive. Simulation results on a speaker independent vowel recognition task are presented to show the improved generalization using the pruning algorithm. *************************************************************************** To retrieve: unix> ftp cheops.cis.ohio-state.edu (or 128.146.8.62) Name: anonymous Password: neuron ftp> cd pub/neuroprose ftp> binary ftp> get sankar.ijcnn91_1.ps.Z ftp> get sankar.ijcnn91_2.ps.Z ftp> quit unix> uncompress sankar.ijcnn*.ps unix> lpr sankar.ijcnn91_1.ps sankar.ijcnn91_2.ps Thanks to Jordan Pollack for making this service available! From peterc at chaos.cs.brandeis.edu Tue Apr 23 22:58:55 1991 From: peterc at chaos.cs.brandeis.edu (Peter Cariani) Date: Tue, 23 Apr 91 22:58:55 edt Subject: THRESHOLDS AND SUPEREXCITABILITY In-Reply-To: Christof Koch's message of Thu, 18 Apr 91 13:59:13 PDT <910418135913.2040bf1d@Iago.Caltech.Edu> Message-ID: <9104240258.AA22066@chaos.cs.brandeis.edu> In reply to Christoff Koch's message: I had thought I had made myself clear regarding time-dependent threshold changes. Raymond and others have found a triphasic threshold time course following an action potential: the threshold initially goes up (refractory phase) far above its resting level, then it goes DOWN below its resting level (superexcitable phase), then it goes back up above its resting level (depression phase) and returns to its resting level. Superexcitability typically begins several milliseconds after the action potential and can last milliseconds to hundreds of milliseconds. While almost every neural model includes the refractory phase, I have yet to see one (outside of Raymond's model for threshold changes) which includes superexcitability. Actually, I had read Christoff Koch's and Tomasso Poggio's otherwise excellent review paper "Biophysics of computation: neurons, synapses and membranes." in Synaptic Function, Edelman, Wall & Cowan eds Wiley, 1989 when I began to look at neural models about a year and a half ago, and there is a section on conduction blocks but no mention of threshold changes or of any of the work by Lettvin & company. I know the book "Methods in Neuronal Modelling" and yes it has many detailed state-of-the-art models, but I explicitly looked there when the book first came out for a model incor- porating nonmonotonic threshold recovery curves and as I recall none of the models in that book even come close. This is not to say they are bad models (on the contrary, they are the best we have), but that there are some phenomena of potential import which are being overlooked. Peter Cariani From ken at cns.caltech.edu Wed Apr 24 19:00:04 1991 From: ken at cns.caltech.edu (ken@cns.caltech.edu) Date: Wed, 24 Apr 91 16:00:04 PDT Subject: 2nd announcement: Methods in Computational Neuroscience course Message-ID: <9104242300.AA16648@cns.caltech.edu> COURSE ANNOUNCEMENT: METHODS IN COMPUTATIONAL NEUROSCIENCE *** 2ND (AND LAST) NOTICE *** APPLICATIONS ARE DUE MAY 15 *** We are writing to bring to your attention the course ``Methods in Computational Neuroscience", to be offered Aug 5 -- 30, 1991 at the Marine Biological Laboratory, Woods Hole, Massachusetts. The course will provide intensive training in methods of neural modeling to 20 graduate students, postdocs, and faculty. Course directors are James Bower, Christof Koch, and Kenneth Miller (Associate Director), from Caltech. The course will include both lectures, and a laboratory in which each student will undertake a computational project of their own choosing. Each student will be provided with a Unix workstation (Sun Sparcstation II) and will be trained in doing neural simulations using the general purpose neural simulator GENESIS that has been developed at Caltech. GENESIS is publicly distributed without cost, so students can bring the skills acquired back to others at their institution. Course lectures are roughly divided into two sections. The first section will focus on biophysically realistic, detailed simulations of activity patterns in neurons and neuronal networks. Methods for simulating the electrical properties of single neurons will be introduced (compartmental models, active currents, interactions between synapses, calcium dynamics). Numerical and mathematical techniques necessary for such simulations will be presented, as will mathematical methods appropriate for understanding simplified network models (e.g. theory of dynamical systems). Applications to networks will be drawn from the invertebrate and vertebrate literature (central pattern generators, visual system of the fly, mammalian hippocampus, olfactory cortex). Faculty for this section will include: Paul Adams, SUNY Stony Brook Christof Koch, Caltech Idan Segev, Hebrew University Matthew Wilson, U. of Arizona Michael Mascagni, Supercomputing Research Center and NIH Jim Bower, Caltech Roger Traub, IBM John Rinzel, NIH Avis Cohen, U. Maryland Nancy Kopell, Boston U. Eve Marder, Brandeis Sylvia Ryckebusch, Caltech The second section will build on the first, focusing on learning, development, and higher functions. Topics such as LTP in hippocampus, development of the visual cortex and of the retino-tectal projection, and motion perception will be presented. Models at both the cellular and network levels will be discussed. More abstract networks, including biologically relevant learning algorithms and connectionist models, will also be examined from a neurobiological point of view. Faculty for this section will include: Charles Stevens, Salk Institute Anthony Zador, Yale Kenneth Miller, Caltech David MacKay, Caltech David Zipser, UCSD Terry Sejnowski, UCSD and the Salk Institute Christof Koch, Caltech Bill Bialek, NEC Corp. Rodney Douglas, Oxford David Van Essen, Caltech Richard Anderson, MIT Ted Adelson, MIT Requests for applications should be sent to: Ms. Florence Dwane Admissions Coordinator Marine Biological Laboratory Woods Hole, MA 02543 Applications should be sent to Ms. Dwane by May 15 1991. Financial aid is available for all students that need it. We believe this course provides a unique opportunity to develop skills in neural modeling. Please bring the course to the attention of your colleagues. We encourage you and your colleagues to consider attending. From lyle at ai.mit.edu Wed Apr 24 12:26:52 1991 From: lyle at ai.mit.edu (Lyle J. Borg-Graham) Date: Wed, 24 Apr 91 12:26:52 EDT Subject: THRESHOLDS AND SUPEREXCITABILITY In-Reply-To: Peter Cariani's message of Tue, 23 Apr 91 22:58:55 edt <9104240258.AA22066@chaos.cs.brandeis.edu> Message-ID: <9104241626.AA28990@wheat-chex> RE: Peter Cariani's message With the ever-increasing types (or as Paul Adams puts it, "Balkanization") of voltage/time/2nd-messenger(e.g. Ca++)-dependent channels being identified in various cells, it is clear that any canonical threshold model per se (tri-phasic time-dependency or otherwise) is problematic. The emerging picture is that there are a broad spectrum of channel combinations which may be participating in different contexts/neurons, resulting in quite different response modes. Whether a given model reproduces that spectrum or not depends, of course, on what the model is claiming to represent. The intricate and subtle interactions between channels in bullfrog slime neurons or hippocampal pyramidal cells (for example) are neat, and may *actually* be important for function. But that does not mean a much simpler membrane model would be more appropriate for a given simulation. Lyle Borg-Graham From koch at CitIago.Bitnet Thu Apr 25 05:51:29 1991 From: koch at CitIago.Bitnet (Christof Koch) Date: Thu, 25 Apr 91 02:51:29 PDT Subject: THRESHOLDS AND SUPEREXCITABILITY In-Reply-To: Your message <9104240258.AA22066@chaos.cs.brandeis.edu> dated 23-Apr-1991 Message-ID: <910425024432.2041a557@Iago.Caltech.Edu> I agree with Peter C. that none of the current models contain a superexcitable phase. However, do we know to what extent this phenomena occurs under physiological circumstances in say, cortical cells. As far as I remember Raymond and Lettvin's J. Physiol. paper on the timecourse of the spiking threshold, they worked on alpha motoneurons and made the cell fire hundreds of time. In a cortical cell, discharge of an action potential will transiently open calcium channels. The potential will return to the resting potential and will be pulled below to more hyperpolarizing levels by the combined action of several voltage and/or calcium-dependent potassium currents, in particular I_C, I_M and I_AHP. They effectively implement a relative refractory period. In hippocampal cells, action potentials are sometimes followed by ADP, i.e. after-depolarizations, caused by charge coming back from the dendrites. I wonder whether anybody knows whether the threshold for action potential initiation goes DOWN during these ADPs. This would be the closest phenomena to "superexcitability" I know of. Christof From lyle at ai.mit.edu Thu Apr 25 12:30:20 1991 From: lyle at ai.mit.edu (Lyle J. Borg-Graham) Date: Thu, 25 Apr 91 12:30:20 EDT Subject: THRESHOLDS AND SUPEREXCITABILITY In-Reply-To: Christof Koch's message of Thu, 25 Apr 91 02:51:29 PDT <910425024432.2041a557@Iago.Caltech.Edu> Message-ID: <9104251630.AA01291@substantia-nigra> > In a cortical cell, discharge of an action potential will transiently open > calcium channels. The potential will return to the resting potential > and will be pulled below to more hyperpolarizing levels by the combined > action of several voltage and/or calcium-dependent potassium currents, > in particular I_C, I_M and I_AHP. They effectively implement a relative > refractory period. > In hippocampal cells, action potentials are sometimes followed by ADP, i.e. > after-depolarizations, caused by charge coming back from the dendrites. Probably the reason a distinct "ADP" has been described in hippocampal pyramidal cells is that there is a ~1msec transient hyperpolarization (fAHP, or fast after-hyperpolarization) during the spike repolarization, due to the fast Ca-dependent K current (I_C). Since this current inactivates (possibly because of voltage) quickly, a subsequent depolarization results. However, the distributed charge of the dendrites which underlies this depolarization is most likely in other cells (the estimation of which is a good candidate for modelling), and thus the ADP phase is there as well. > I wonder whether anybody knows whether the threshold for action potential > initiation goes DOWN during these ADPs. I suppose it depends on what is meant by threshold, i.e. current or voltage. If the Na voltage 'threshold' doesn't change [which it does] then the ADP would lower the current threshold since, of course, the dendritic charge is lending a hand. > This would be the closest phenomena to "superexcitability" I know > of. For what it is worth, it might be better to reserve the term "excitability" for the state of intrinsic membrane properties. As far as "superexcitability" goes, perhaps the I_A and I_D K+ currents in hipppocampus may participate. The inactivation range of these currents include voltages below spike threshold, with varying time constants (I_A on the msec scale, I_D on the 10 to 100 msec scale). Thus spike activity tends to shut them off, which will contribute to excitability. Now the interesting thing to do is to identify the physiological pathways that modulate these guys. Lyle From SAYEGH at CVAX.IPFW.INDIANA.EDU Thu Apr 25 22:54:32 1991 From: SAYEGH at CVAX.IPFW.INDIANA.EDU (SAYEGH@CVAX.IPFW.INDIANA.EDU) Date: Thu, 25 Apr 1991 22:54:32 EDT Subject: proceedings 3rd NN & PDP + Air Mail Postage Message-ID: <910425225432.20202a97@CVAX.IPFW.INDIANA.EDU> In the announcement of the proceedings of the third conference on Neural Networks and Parallel Distributed Processing at Indiana-Purdue University, there was a minor mix-up in the list of papers. We also received a number of requests from Europe and Japan inquiring about Air Mail costs. Here is the info: Proceedings can be obtained by writing to: Ms. Sandra Fisher Physics Department Indiana University-Purdue University Ft Wayne, IN 46805 and including $5 + mailing and handling costs as follows $1 for the US, $2 for Canada and Mexico, $8.50 for all others by Air Mail or $3.95 by surface mail. Checks should be made payable to The Indiana-Purdue Foundation. The 109 page proceedings contain the following papers: INTEGRATED AUTONOMOUS NAVIGATION BY ADAPTIVE NEURAL NETWORKS Dean A. Pomerleau Department of Computer Science Carnegie Mellon University APPLYING A HOPFIELD-STYLE NETWORK TO DEGRADED PRINTED TEXT RESTORATION Arun Jagota Department of Computer Science State University of New York at Buffalo RECENT STUDIES WITH PARALLEL, SELF-ORGANIZ- ING, HIERARCHICAL NEURAL NETWORKS O.K. Ersoy & D. Hong School of Electrical Engineering Purdue University INEQUALITIES, PERCEPTRONS AND ROBOTIC PATH- PLANNING Samir I. Sayegh Department of Physics Indiana University-Purdue University GENETIC ALGORITHMS FOR FEATURE SELECTION FOR COUNTERPROPAGATION NETWORKS F.Z. Brill & W.N. Martin Department of Computer Science University of Virginia MULTI-SCALE VISION-BASED NAVIGATION ON DIS- TRIBUTED-MEMORY MIMD COMPUTERS A.W. Ho & G.C. Fox Caltech Concurrent Computation Program California Institute of Technology A NEURAL NETWORK WHICH ENABLES SPECIFICATION OF PRODUCTION RULES N. Liu & K.J. Cios The University of Toledo PIECE-WISE LINEAR ESTIMATION OF MECHANICAL PROPERTIES OF MATERIALS WITH NEURAL NETWORKS I.H. Shin, K.J. Cios, A. Vary* & H.E. Kautz* The University of Toledo & NASA Lewis Re- search Center* INFLUENCE OF THE COLUMN STRUCTURE ON INTRA- CORTICAL LONG RANGE INTERACTIONS E. Niebur & F. Worgotter California Institute of Technology LEARNING BY GRADIENT DESCENT IN FUNCTION SPACE Ganesh Mani University of Wisconsin-Madison SUCCESSIVE REFINEMENT OF MULTI-RESOLUTION REPRESENTATIONS OF THE ENVIRONMENT IN CONNECTIONIST NETWORKS Vasant Honovar and Leonard Uhr Computer Sciences Department University of Wisconsin-Madison A NEURAL ARCHITECTURE FOR COGNITIVE MAPS Martin Sonntag Cognitive Science & Machine Intelligence Lab University of Michigan From ken at cns.caltech.edu Fri Apr 26 10:20:41 1991 From: ken at cns.caltech.edu (ken@cns.caltech.edu) Date: Fri, 26 Apr 91 07:20:41 PDT Subject: Network: A New, Affordable Subscription Price Message-ID: <9104261420.AA18436@cns.caltech.edu> One of the several good neural network journals to have started in the last year or two is Network: Computation in Neural Systems, edited by Daniel Amit. The journal is self-consciously interdisciplinary between Neuroscience and physical/mathematical/computational sciences. The scope of the articles is similar to that of the articles found in Neural Computation, but Network includes primarily full papers (long articles) whereas Neural Computation includes primarily letters (short articles). My own interest in Network is motivated in no small part by the fact that the long, complete version of the paper of David MacKay and myself analyzing Linsker's simulations is found in Network, Vol. 1, #3; a short version was in Neural Computation. To mention a very few among many other recent authors: D. Willshaw, J. Hopfield, M. Abeles, P. Simic, D. Parisi, D. Amit, A. Selverston, G. Toulouse, V. Braitenberg, J. Nadal, E. Rolls. Network has been relatively inaccessible, at least in the US. I think this has been largely due to its very high subscription price. So I am writing to advertise the fact that Network has just instituted a very affordable subscription price for individuals. The new price for individuals is $37.90/year (US, Canada and Mexico), 17.30 lbs/year (UK), or 20.50 lbs/year (other). This price applies to Volume 1 (1990), as well as to the current volume, Vol. 2 (1991). To order: US/Can/Mex: American Institute of Physics, Subscriber Services, 500 Sunnyside Blvd, Woodbury, NY 11797-2999; 516-349-7800 x628; Other: Journals Mkting Dept., IOP Publishing, Techno House, Redcliffe Way, Bristol BS1 6NX, England. I won't spend bandwidth listing contents of recent issues; people at these addresses can send you literature. Ken Miller From ken at cns.caltech.edu Fri Apr 26 11:39:35 1991 From: ken at cns.caltech.edu (ken@cns.caltech.edu) Date: Fri, 26 Apr 91 08:39:35 PDT Subject: application material for Methods in Computational Neuroscience course Message-ID: <9104261539.AA18493@cns.caltech.edu> With respect to the previous 2nd announcement about the Methods of Computational Neuroscience course: I have received numerous inquiries as to how to obtain application materials rapidly (since the deadline for receipt of applications is soon -- May 15), so it seems merited to post the following info to the net as a whole: --- To repeat, the person who must be contacted to obtain application material, and to whom completed applications should be sent, is: Ms. Florence Dwayne, Admissions Coordinator, Marine Biological Laboratory, Woods Hole, MA 02543. [Note, I left the 'y' out of her last name in previous posting]. --- THERE IS NO E-MAIL ADDRESS FOR OBTAINING OR RETURNING APPLICATIONS. --- Ms. Dwayne can be reached by phone at 508-548-3705, extension 216. --- IF NECESSARY she can be reached by fax at 508-457-1924. The MBL fax machine is extremely busy, so they prefer not to send applications out by fax. However they will do so if necessary, particularly for overseas applicants. --- If completed applications must be returned quickly, please if possible use express mail or a courier service (i.e. Federal Express, DHL). Again, applications can IF NECESSARY be returned by fax, but this is strongly discouraged if an alternative is available. Ken Miller From worth at park.bu.edu Fri Apr 26 12:37:23 1991 From: worth at park.bu.edu (Andrew J. Worth) Date: Fri, 26 Apr 91 12:37:23 -0400 Subject: IJCNN-91-SEATTLE call for volunteers Message-ID: <9104261637.AA01878@park.bu.edu> --------------------------------------------------------------------------- IJCNN-91-Seattle Call for Volunteers July 8-12th, 1991 Seattle, Washington, USA. --------------------------------------------------------------------------- The International Joint Conference on Neural Networks (IJCNN-91-SEATTLE) has volunteer positions available. If you or anyone you know would like to exchange admittance to the conference for working as a volunteer, please respond directly to me at the e-mail address below. In the past, volunteers have given approximately 20 hours of labor (spread out over the entire conference) to receive: o admittance to the conference o a full set of proceedings o attendance to a limited number of tutorials (while working) The exact benefits are still being worked out. Volunteer positions include helping at: o Stuffing Conference Proceedings o Poster Sessions o Technical Sessions o Evening Plenary Sessions o Social Events o OPTIONAL duty: Tutorials If you are interested in volunteering, please respond directly to me with the following information: o Electronic Mail Address o Last Name, First Name o Address o Country o Phone number o Volunteer Position Preference Positions will be filled on a first commit first served basis. There will be no funding available for volunteer's travel and lodging expenses. PLEASE RESPOND TO: worth at park.bu.edu Thank you, Andy. ------------------------------------------------------------------------- Andrew J. Worth Cognitive & Neural Systems Dept. IJCNN-91 Volunteer Chair Boston University worth at park.bu.edu 111 Cummington Street, Rm 244 (617) 353-6741 Boston, MA 02215 ------------------------------------------------------------------------- From whsu at ecn.purdue.edu Sat Apr 27 12:02:25 1991 From: whsu at ecn.purdue.edu (William Hsu) Date: Sat, 27 Apr 91 11:02:25 -0500 Subject: Training standard one hideen layer backprop for Mglass Data Message-ID: <9104271602.AA10221@panther.ecn.purdue.edu> Dear Researchers, I have been trying to train a standard backprop network with one hidden layer with varying number of sigmoid units to learn the Mglass (Chaotic Data ref. Platt, Moody) Time series. I have not been able to acheive Normalized root mean square values of < 0.4 for archiectures to 50 hidden units. The architectures I have looked at includes 4-N-1. for various N. where the output unit is a linear summation unit (without sigmoid). I would appreciate hearing from people who have had similar experience or better still, people who have successful experience with the above setup. Pls email directly to me. Thank you very much. From booker at starbase.MITRE.ORG Mon Apr 29 14:51:13 1991 From: booker at starbase.MITRE.ORG (Lashon Booker) Date: Mon, 29 Apr 91 14:51:13 EDT Subject: ICGA-91 Preliminary Program / Call for Participation Message-ID: <9104291851.AA11805@starbase.mitre.org> Fourth International Conference on Genetic Algorithms ICGA91 13-16 July, 1991 Univ. California San Diego La Jolla, CA PRELIMINARY PROGRAM The Fourth International Conference on Genetic Algorithms (ICGA-91), will be held July 13-16, 1991 at the University of California - San Diego in La Jolla, CA. This meeting brings together an international community of scientists from academia and industry interested in algorithms suggested by the evolutionary process of natural selection. Topics of particular interest include: Holland's Genetic Algorithm and Classifier Systems, machine learning and optimization using these systems, and their relations to other learning paradigms (e.g., connectionist networks). Papers discussing how genetic algorithms and classifier systems are related to biological modeling issues (e.g., evolution of nervous systems, computational ethology, artificial life) have been encouraged. Interest in Genetic Algorithms (GAs) has risen steadily, and ICGA-91 drew more than 160 submissions. The conference committee believes it is especially important at this time of rapid growth in GA research to create a core of common understanding that will ensure future communication among the GA community. It is for this reason that ICGA-91 will mix invited talks and a small number of plenary presentations with as much small-group interaction as possible. Poster sessions will be the default mode for papers; some of the strongest papers to be included in ICGA-91 will be presented as posters. There will be no parallel sessions. CONFERENCE CO-CHAIRS: Kenneth A. De Jong J. David Schaffer Vice Chair: David E. Goldberg Program Co-Chairs: Richard K. Belew Lashon B. Booker Publicity Chair: David E. Goldberg Financial Chair: Gilbert Syswerda Local Arrangements: Richard K. Belew PROGRAM COMMITTEE: Emile Aarts, NV Phillips Richard K. Belew, UCSD Lashon B. Booker, MITRE Yuval Davidor, Weizmann Inst. L. David Davis, TICA Associates Kenneth A. De Jong, George Mason Univ. Larry Eschelman, Phillips Labs Stephanie Forrest, Univ. New Mexico John Greffenstette, Naval Research Lab. David E. Goldberg, Univ. Illinois Paulien Hogeweg, Univ. Utrecht John H. Holland, Univ. Michigan Gunar Liepins, Oakridge Natl. Lab. Heinz Muehlenbein, GMD, Germany John Koza, Stanford University Gregory Rawlins, Univ. Indiana Rick Riolo, Univ. Michigan George Robertson, Xerox PARC J. David Schaffer, Philips Labs Steve F. Smith, Carnegie-Mellon Univ. Gilbert Syswerda, BBN Tom Westerdale, Univ. London Stewart Wilson, Roland Inst. Darrell Whitley, Colorado State Univ. SUPPORTED BY: International Society for Genetic Algorithms Office of Naval Research Naval Research Laboratory TENTATIVE SCHEDULE Sat Sun Mon Tues AM 1: Tutorials Plenary 1 Plenary 3 Plenary 5 AM 2: Tutorials Poster 1 Poster 2 Poster 3 Lunch Lunch Lunch Lunch PM 1: Invited Talks Plenary 2 Plenary 4 Panel PM 2: Invited Talks Workshops Workshops Bus. Mtg. Evening: Social Mixer Workshops Banquet The schedule has been arranged so that it will be possible for participants to fly in Saturday (13 July) morning and leave Tuesday (16 July). We expect, however, that most ICGA-91 participants will want to attend the tutorials Saturday morning and the panel session and perhaps the ISGA business meeting Tuesday afternoon. Also, space is available for workshops after the business meeting and perhaps on Wednesday. Housing for these additional nights can be easily reserved at the same per-night rate; specify these on the registration form. INVITED SPEAKERS The ICGA-91 conference will officially begin on Saturday afternoon, 13 July with a series of invited presentations by leading researchers from computer science, mathematical biology, economics and philosophy: John Holland (University of Michigan) "Complex Adaptive Systems" Marcus Feldman (Stanford University) "Optimality and the Evolution of Recombination" John Miller (Santa Fe Institute) "Artificial Adaptive Agents in Economics" William Wimsatt (University of Chicago) "Developmental constraints on evolving systems" After these individual presentations and a dinner break, the guests will participate in a panel discussion delineating important research questions for the rest of the ICGA-91 meeting. TUTORIALS On Saturday morning, before the conference officially begins, we will be offering a series of tutorials. Two "tracks" have been designed, with one for ICGA-91 attendees who are new to Genetic Algorithms and one for attendees interested in advanced topics. An additional fee of $60 ($25 for students) will be charged to attend any and all tutorials; this fee includes course materials, lunch and the coffee break. Track I I.A Introduction to Genetic Algorithms (John Greffenstette) I.B Applications of Genetic Algorithms (L. David Davis) I.C Introduction to Classifier Systems (Rick Riolo) Track II II.A Genetic Algorithms Theory (Dave Goldberg and Gunar Liepins) II.B Biological Insights into GAs (Rik Belew and Chuck Talyor) PRE-REGISTRATION FOR THESE TUTORIALS (ON THE ENCLOSED REGISTRATION FORM) IS STRONGLY ADVISED, AS SPACE IS LIMITED. WORKSHOPS With the growing general interest in Genetic Algorithms (GAs), an important role for special interest groups is emerging. ICGA-91 has reserved time and space for smaller groups of people to meet and discuss some of these specialized aspects of GA research. Many of these topics were organized in response to large numbers of submissions in an area, many of which could not be published as part of the ICGA-91 Proceedings. If your work is related to one of these workshop topics, we encourage you to contact the chairperson of the relevant workshop. Also, if you have ideas for other workshops and would be willing to act as an organizer, contact Rik Belew (rik at cs.ucsd.edu) with your suggestion and we will do our best to publicize it. Here are the workshops presently organized: I. Machine Learning with GAs John Greffenstette (gref at aic.nrl.navy.mil) How well do GAs apply traditional machine learning problems, such as concept learning from examples, conceptual clustering, etc. How do GAs and classifier systems relate to other methods of reinforcement learning (e.g., Q-learning)? II. Evolutionary Strategie Hans-Paul Schwefel (schwefel at lumpi.informatik.uni-dortmund.de) A great deal of research, pioneered by Rechenberg, Schwefel and others, has been done in Germany on a class of "evolutionary strategies" closely related to Holland's form of the GA. This workshop is designed to acquaint members of the GA community unfamiliar with the theory and application of these techniques. III. Parallel Implementations of the GA Heinz Muhlenbein (muehlen at gmdzi.uucp) John McInerney (john at cs.ucsd.edu) GAs are proving themselves to be an important programming methodology for distributed and massively parallel machines. Moreover, this workshop will survey existing implementations of the GA on parallel machines and consider outstanding research issues. IV. GAs and Connectionist Networks Richard K. Belew (rik at cs.ucsd.edu) Both GAs and connectionist ("neural") networks represent active new approaches to machine learning. This workshop will consider hybrids of these methods, as well as attempts to relate GAs and connectionist networks as biological models of evolution and learning, resp. V. GAs and Artificial Life Peter Todd (todd at psych.stanford.edu) GAs are becoming central ingredients in a new class of "artificial life" simulations. This workshop will acquaint participants with some of the most important ALife experiments to date, and consider the special demands it makes on GAs. VI. Software support and test suites for GAs L. David Davis (70461.1552 at compuserve.com) Nici Schraudolf (nici at cs.ucsd.edu) Several large simulators have been developed for GAs, and a number of fitness function test suites have also been proposed. Both advances promise to make the job of the GA investigator easier, and also facilitate comparison across investigators. This workshop will survey these alternatives and attempt to form a consensus for future developments. VII. Biological influences in problem solving with Genetic Algorithms Alan Schultz (schultz at aic.nrl.navy.mil) The purpose of this workshop is to bring together GA researchers who are interested in improving the GA's performance in problem solving with mechanisms that are analogical to, or inspired by biological mechanisms. VIII. Theory of Genetic Operators Nick Radcliffe (njr at castle.edinburgh.ac.uk) There have been a number of recent developments which extend the applicability of schema analysis and our understanding of intrinsic parallelism. This workshop will examine these and look at new operators which have theoretical underpinning. TRAVEL INFORMATION Please make your travel arrangements early. Summer is a busy travel period in Southern California. San Diego can be reached conveniently by air, train or bus from Los Angeles. There are many direct flights to San Diego from around the United States. A special conference fare has been negotiated with American and Northwest Airlines for the ICGA conference. This applies to domestic and Canadian reservations only. A 40% discount will be offered for any roundtrip coach fare type, while 5% discount will be offered on the lowest available fare type, subject to capacity control, purchase time limits and other applicable fare rules. American Airlines: File Number 06718G; phone number: 1-800-433-1790. Northwest Airlines: File Number 04664; phone number: 1-800-328-1111. If you are arriving by air to the San Diego Airport (Lindbergh Field) look for a Peerless Shuttle; we have negotiated special rates for transportation to campus. Your registration confirmation package will include a flyer for your use. The rate is $14 for one person and $5 per additional person. A cab fare from the airport to UCSD is about $30. Ask the driver to take you to Warren College Conference Desk on Warren Campus. (You will probably need to show the campus map to the driver in order to find the registration desk, although he will know where UCSD is!) Housing registration can be completed at Warren College Conference Desk. Meeting registration will take place at La Casa during the Welcome Reception Friday evening or at the Price Center's Theatre Lobby on Saturday, Sunday or Monday. PARKING The distances on campus are not large and a car is not necessary. For example, the walk between the housing and the meeting facilities is about 3 minutes. However, if you are arriving by private or rental car, you will require a parking permit. These may be purchased at the Registration desk for $15.00 for a week or $4.00 per day. Parking permits are required seven days a week from 7am to 11pm. Monitoring of permits tends to be quite diligent. There is a parking moratorium on Saturday July 13 in parking lot 503 on Warren Campus; so, feel free to park anywhere there while you register. WEATHER The weather in La Jolla in July is usually very pleasant with temperatures around 25 degrees C (77 degrees F) and low humidity. Evenings can be cooler, so a light jacket or sweater may be useful. Occasionally very hot, dry weather (Santa Ana) can occur for a few days if the prevailing winds shift and come from the inland deserts. The Pacific Ocean is warm enough for swimming. We suggest you bring along swim-wear and a beach towel. REGISTRATION AND INFORMATION DESKS The registration and information desks will be open as follows: Conference registration will be in the Price Center Theatre Lobby: 7/12 - 7/15 7:00am - 5:00pm 7/16 7:00am - 2:00pm Housing registration only will be at the Warren College Desk, Daily 8:00am - 12:00 midnight If you need further assistance with some practical matter at any time, please see a member of the Organizing Committee. MESSAGES A telephone message service for incoming calls to conference participants will be set up. Please check the board in the foyer of the theater for messages. The conference telephone number will be communicated later. Outgoing calls may be made from public telephones located throughout the campus. If you have special problems with communication to the outside world during the conference, please stop by the registration desk or ask a member of the Organizing Committee. We are making arrangements to allow INTERNET access from terminals near the conference headquarters. REGISTRATION FEES The registration fee for conference participants includes entry to all technical sessions, conference proceedings, refreshment breaks, the receptions Friday and Saturday evenings, breakfast and lunch Sunday through Tuesday. The following fees have been established (all figures in U.S. dollars): Before After 6/1/91 6/1/91 Conference Participant $225 $275 Student $100 $125 A limited fund has been set aside to assist students and scientists with their travel expenses. If you are interested in obtaining such assistance and have no other form of travel support, send a letter describing your situation and needs to: Gilbert Syswerda International Society for Genetic Algorithms 53 Lake Street Winchester, MA 01890 (syswerda at bbn.com) Students should have their advisor certify their student status and that other funds are not available. REMITTANCE PROCEDURES Please fill out the registration form found elsewhere in this booklet and return with payment. Payment must be in U.S. dollars drawn on a U.S. bank. Checks should be made out to UC Regents, I.C.G.A. Personal checks, cashier's checks or international money orders in U.S. dollars are acceptable. Do not send cash. Cash (again in U.S. dollars) may be used for payment at the registration desk. Please note that credit cards cannot be accepted for payment under any circumstances. Do not send payment that is not in U.S. dollars. Registration form and payment should be sent to: Lene Hartman Conference Manager, I.C.G.A. UCSD 9500 Gilman Drive La Jolla, CA 92093-0513, USA The pre-registration deadline is June 1, 1991. Participants may register at a later time, including at the conference itself, although we strongly recommend that you register in advance. Mid-July still seems quite far away, but the program committee is already very excited by the prospect of a strong meeting of the GA community. Be sure to get your registration forms in early as attendance promises to be high and space is limited. Once the registration form and remittance have been received, the conference manager will send a receipt and directions to the participant's address. ON-CAMPUS HOUSING AND MEAL REGISTRATION INFORMATION Campus housing is available to conference participants as a package plan only - three nights' housing from Saturday through Monday. This housing is in two-bedroom apartments with shared living room, kitchen and bathroom. Each bedroom has two single beds and is rented as either a single (if only one person occupies the room) or as a double (if two people occupy the room). If you have made arrangements to share a double room with another I.C.G.A. participant, please include the name on the registration form. If you are willing to share an apartment but have not made such arrangements, we will coordinate such requests and assign you to a (same sex) apartment. Conference participants and accompanying guests may stay in campus housing several nights before or after the conference. Room rates for these additional nights are noted on the registration form. The rates given are per person, per night except for the family unit which is a per apartment, per night rate. Also make sure you specify your arrival and departure dates. A special lunch only package is available for accompanying guests. The cost of this package is $39 if purchased in advance; $64 if purchased after June 1. This purchase enables the accompanying guests to eat with the participants in the Price Center Ballroom. Accompanying guests choosing not to purchase meals as part of a package plan may purchase meals at one of the five restaurants within the Price Center or at one of the other Food Service facilities throughout campus. Additional tickets for the Receptions and the Banquet must be purchased in advance. BANQUET On Monday, 15 July a banquet will be held at Torrey Pines Inn. This hotel is perched overlooking the Pacific Ocean, and adjacent to the world famous Torrey Pines Golf course. The evening will begin with drinks around the Inn's Olympic swimming pool, watching the sun set into the Pacific. We will then be served a choice of beef, chicken or vegetarian dinners. (Please indicate your choice on the registration form.) Because we want to encourage attendance by the families of conference participants, we have made arrangements for a special children's room and dinner. Children will be served a choice of hamburger or vegetarian pizza, and supervised and entertained in a room of their own. BANQUET RESERVATIONS MUST BE MADE AS PART OF PRE-REGISTRATION. NO BANQUET TICKETS WILL BE AVAILABLE FOR SALE AT THE CONFERENCE. HOTEL ACCOMMODATIONS We have reserved a small number of rooms in the three hotels listed below: Torrey Pines Inn 11480 North Torrey Pines Road, La Jolla, CA 92037 Located next to the world-class Torrey Pines Golf Course and offers a quiet, rustic atmosphere. The distance between the Inn and UCSD is 1 1/2 miles and is a very pleasant walk. The Inn offers transportation to and from campus. The rate is $70 for double and $65 for single occupancy. Golf facilities and an olympic- size pool for lap swimming are available. Complimentary cocktail and newspaper. Please call (619) 453- 4420 for a reservation. La Jolla Village Inn, 3299 Holiday Court, La Jolla, CA 92037 Located 1/2 mile from UCSD this hotel offers complimentary transportation to and from the airport and the campus. The rate is $69 for double and $64 for single occupancy. Please call 1-800 854-2900 for a reservation. Sheraton Grande 10950 North Torrey Pines Rd., La Jolla, CA 92037 This new luxury hotel offers a fantastic view from each room of either the Pacific Ocean or the beautiful Torrey Pines Golf Course. Located atop Torrey Pines, it is convenient to UCSD campus and minutes from La Jolla. This exciting property is complete with several restaurants, cocktail lounges - all offering sunset views. Tennis, golf, swimming and a health spa are just a few of the hotel's amenities. The rate is $85 for either single or double. Please call (619) 558-1500 for a reservation. ON-CAMPUS RECREATION The campus housing is located near the recreation center which offers an olympic-size pool, racquetball and volleyball courts. Tennis courts are located near the housing complex (please bring your own racket and balls). Recreation cards may be purchased for a small fee ($14 single/$26 family) at the registration desk. PROGRAM FOR ACCOMPANYING GUESTS Transportation to Balboa Park, the San Diego Zoo, the Wild Animal Park and Old Town can be arranged. A modest fee will be charged. Sign-up sheets will be available at registration. The tours will be Sunday, Monday and Tuesday. Please sign up one day before your expected participation. GOLF The Torrey Pines Golf course is 1 1/2 miles from campus and the rates are: $35 per week-day, cash only $40 per week-end day $22 per cart $20 for clubs Reservations for a business day may be made two business days in advance at 1pm and for a week- end day five business days in advance. So, if you want to play Sunday, July 14 you must call Monday, July 8 at 8am. Call (619) 453-0380. INSURANCE Foreign participants are strongly encouraged to take out a traveller's health insurance policy before arriving in the United States. SMOKING UCSD has been designated by the university chancellor as a non-smoking environment. Smoking is not permitted within any public buildings on the campus. However, smoking is permitted in your apartment and on the grounds. SYNOPSIS OF IMPORTANT DATES May 1, 1991: Revised, camera-ready paper due June 1, 1991: Pre-registration deadline for conference and on-campus housing June 15, 1991: Audio visual and software demonstration requirements must be received July 13-16, 1991: Registration, conference starts ICGA91 - THE FOURTH INTERNATIONAL CONFERENCE ON GENETIC ALGORITHMS REGISTRATION FORM Dr./Mr./Ms.// Family Name:__________________ First Name: _________ Title: ___________________________ Institution: ____________________ Mailing Address: ___________________________________________________ Telephone: ______________________ FAX or Telex: ___________________ Electronic Mail Address: ___________________________________________ Name of Accompanying Person: _______________________________________ Preferred Roommate (for those choosing double rooms) _______________ Arrival Date: ____________________ Departure Date: _________________ ================================================================ CONFERENCE REGISTRATION Before On or After Quantity 6/1/91 6/1/91 Participant Registration ____ US $225 US $275 ________ Student Registration ____ US $100 US $125 ________ T-Shirt Size S ___ M ___ L ___ XL ___ Tutorial 7/13 Participant ____ US $ 60 US $ 85 ________ (Lunch included) Tutorial 7/13 Student ____ US $ 25 US $ 50 ________ (Lunch included) Track I ____ Track II ____ Registration sub-total ________ ================================================================ MEALS, RECEPTION, BANQUET Before On or After Quantity 6/1/91 6/1/91 Additional Welcome Reception Tickets Friday ____ US $10 US $10 ________ Additional Reception Tickets Saturday ____ US $10 US $10 ________ Lunch Only Package - 3 days (Accompanying Persons Only)____ US $39 US $64 ________ Banquet Ticket Adult ____ US $30 US $30 ________ Banquet Ticket Child (under 12) ____ US $ 8 US $ 8 ________ Meals sub-total ________ Banquet Choices: Adult: Beef ___ Chicken ___ Vegetarian ___ Child: Hamburger ___ Vegetarian ___ Pizza ___ PLEASE PURCHASE TICKET NOW IF YOU PLAN TO ATTEND. YOU WILL BE UNABLE TO BUY IT UPON ARRIVAL. ======================================================== APARTMENT HOUSING REGISTRATION Quantity Saturday through Monday (3 nights) Single Room $129/person ____ _________ Double Room $105/person ____ _________ Family(*) Unit $186/family ____ _________ Addl Single Nights $43/person/nite ____ _________ Addl Double Nights $35/person/nite ____ _________ Addl Family(*) Unit Nights $62/apt/nite ____ _________ (*) Family = 2 Adults and at least 1 child Housing sub-total _________ (Registrastion, Food & Housing) TOTAL _________ Payment by: U.S. Check ___ International Money Order ___ Cashier Check ___ Mail completed registration form and check made payable to "UC Regents" to: Lene Hartman, Conference Manager I.C.G.A. UCSD 9500 Gilman Drive La Jolla, CA 92093-0513 U.S.A. ======================================================== (for office use only) Check # _______ Date Rec'd _________ Balance Due _________ Initials ______ From schmidhu at informatik.tu-muenchen.dbp.de Tue Apr 30 03:17:00 1991 From: schmidhu at informatik.tu-muenchen.dbp.de (Juergen Schmidhuber) Date: 30 Apr 91 9:17 +0200 Subject: preprints and reports Message-ID: <9104300717.AA03329(a)kiss.informatik.tu-muenchen.de> Recent preprints and technical reports are available via ftp: ------------------------------------------------------------------ ADAPTIVE DECOMPOSITION OF TIME Juergen Schmidhuber, TUM (Talk at ICANN'91, Helsinki, June 24-28, 1991) In this paper we introduce design principles for unsupervised detection of regularities (like causal relationships) in temporal sequences. One basic idea is to train an adaptive predictor module to predict future events from past events, and to train an additional confidence module to model the reliability of the predictor's predictions. We select system states at those points in time where there are changes in prediction reliability, and use them recursively as inputs for higher-level predictors. This can be beneficial for `adaptive sub-goal generation' as well as for `conventional' goal-directed (supervised and reinforcement) learning: Systems based on these design principles were successfully tested on tasks where conventional training algorithms for recurrent nets fail. Finally we describe the principles of the first neural sequence `chunker' which collapses a self-organizing multi-level predictor hierarchy into a single recurrent network. LEARNING TO GENERATE SUBGOALS FOR ACTION SEQUENCES Juergen Schmidhuber, TUM (Talk at ICANN'91) This paper extends the technical report FKI-129-90 (`Toward compositional learning with neural networks'). USING ADAPTIVE SEQUENTIAL NEUROCONTROL FOR EFFICIENT LEARNING OF TRANSLATION AND ROTATION INVARIANCE Juergen Schmidhuber and Rudolf Huber, TUM (Talk at ICANN'91) This paper is based on FKI-128-90 (announced earlier). ------------------------------------------------------------------- LEARNING TO CONTROL FAST-WEIGHT MEMORIES: AN ALTERNATIVE TO DYNAMIC RECURRENT NETWORKS Juergen Schmidhuber, TUM Technical report FKI-147-91, March 26, 1991 Previous algorithms for supervised sequence learning are based on dynamic recurrent networks. This paper describes alternative gradient-based systems consisting of two feed-forward nets which learn to deal with temporal sequences by using fast weights: The first net learns to produce context dependent weight changes for the second net whose weights may vary very quickly. One advantage of the method over the more conventional recurrent net algorithms is the following: It does not necessarily occupy full-fledged units (experiencing some sort of feedback) for storing information over time. A simple weight may be sufficient for storing temporal information. Since with most networks there are many more weights than units, this property represents a potential for storage efficiency. Various learning methods are derived. Two experiments with unknown time delays illustrate the approach. One experiment shows how the system can be used for adaptive temporary variable binding. NEURAL SEQUENCE CHUNKERS Juergen Schmidhuber, TUM Technical report FKI-148-91, April 26, 1991 This paper addresses the problem of meaningful hierarchical adaptive decomposition of temporal sequences. This problem is relevant for time-series analysis as well as for goal-directed learning. The first neural systems for recursively chunking sequences are described. These systems are based on a principle called the `principle of history compression'. This principle essentially says: As long as a predictor is able to predict future environmental inputs from previous ones, no additional knowledge can be obtained by observing these inputs in reality. Only unpredicted inputs deserve attention. The focus is on a 2-network system which tries to collapse a self-organizing multi-level predictor hierarchy into a single recurrent network (the automatizer). The basic idea is to feed everything that was not expected by the automatizer into a `higher-level' recurrent net (the chunker). Since the expected things can be derived from the unexpected things by the automatizer, the chunker is fed with a reduced description of the input history. The chunker has a comparatively easy job in finding possibilities for additional reductions, since it works on a slower time scale and receives less inputs than the automatizer. Useful internal representations of the chunker in turn are taught to the automatizer. This leads to even more reduced input descriptions for the chunker, and so on. Experimentally it is shown that the system can be superior to conventional training algorithms for recurrent nets: It may require fewer computations per time step, and in addition it may require fewer training sequences. A possible extension for reinforcement learning and adaptive control is mentioned. An analogy is drawn between the behavior of the chunking system and the apparent behavior of humans. ADAPTIVE CONFIDENCE AND ADAPTIVE CURIOSITY Juergen Schmidhuber Technical Report FKI-149-91, April, 26, 1991 Much of the recent research on adaptive neuro-control and reinforcement learning focusses on systems with adaptive `world models'. Previous approaches, however, do not address the problem of modelling the reliability of the world model's predictions in uncertain environments. Furthermore, with previous approaches usually some ad-hoc method (like random search) is used to train the world model to predict future environmental inputs from previous inputs and control outputs of the system. This paper introduces ways for modelling the reliability of the outputs of adaptive world models, and it describes more sophisticated and sometimes much more efficient methods for their adaptive construction by on-line state space exploration: For instance, a 4-network reinforcement learning system is described which tries to maximize the future expectation of the temporal derivative of the adaptive assumed reliability of future predictions. The system is `curious' in the sense that it actively tries to provoke situations for which it {\em learned to expect to learn} something about the environment. In a very limited sense the system learns how to learn. An experiment with a simple non-deterministic environment demonstrates that the method can be clearly faster than the conventional model-building strategy. ------------------------------------------------------------------------ To obtain copies of the papers, do: unix> ftp 131.159.8.35 Name: anonymous Password: your name, please ftp> binary ftp> cd pub/fki ftp> get .ps.Z ftp> bye unix> uncompress .ps.Z unix> lpr .ps Here stands for any of the following six possibilities: icanndec (Adaptive Decomposition of Time) icannsub (Subgoal-Generator). This paper contains 5 partly hand-drawn figures which are not retrievable. Sorry. icanninv (Sequential Neuro-Control). fki147 (Fast Weights) fki148 (Sequence Chunkers) fki149 (Adaptive Curiosity) Please do not forget to leave your name. This will allow us to save paper if you are on our hardcopy mailing list. NOTE: icanninv.ps, fki148.ps, and fki149.ps are designed for European A4 paper format (20.9cm x 29.6cm). ------------------------------------------------------------------------ In case of ftp-problems contact Juergen Schmidhuber Institut fuer Informatik, Technische Universitaet Muenchen Arcisstr. 21 8000 Muenchen 2 GERMANY or send email to schmidhu at informatik.tu-muenchen.de DO NOT USE REPLY! From dfausett at zach.fit.edu Tue Apr 30 13:17:52 1991 From: dfausett at zach.fit.edu ( Donald W. Fausett) Date: Tue, 30 Apr 91 13:17:52 -0400 Subject: preprints and reports Message-ID: <9104301717.AA28357@zach.fit.edu> Greg: Thanks for the additional information. -- Don Fausett From well!moritz at apple.com Tue Apr 30 21:22:58 1991 From: well!moritz at apple.com (Elan Moritz) Date: Tue, 30 Apr 91 18:22:58 pdt Subject: Journal of Ideas, Vol 2 #1 Abstracts Message-ID: <9105010122.AA26275@well.sf.ca.us> +=++=++=++=++=++=++=++=++=++=++=++=++=++=++=+ please post & circulate Announcement ......... Abstracts of papers appearing in Volume 2 # 1 of the Journal of Ideas THOUGHT CONTAGION AS ABSTRACT EVOLUTION Aaron Lynch Abstract: Memory abstractions, or mnemons, form the basis of a memetic evolution theory where generalized self-replicating ideas give rise to thought contagion. A framework is presented for describing mnemon propagation, combination, and competition. It is observed that the transition from individual level considerations to population level considerations can act to cancel individual variations and may result in population behaviors. Equations for population memetics are presented for the case of two-idea interactions. It is argued that creativity via innovation of ideas is a population phenomena. Keywords: mnemon, meme, evolution, replication, idea, psychology, equation. ................... CULTURE AS A SEMANTIC FRACTAL: Sociobiology and Thick Description Charles J. Lumsden Department of Medicine, University of Toronto Toronto, Ontario, Canada M5S 1A8 Abstract: This report considers the problem of modeling culture as a thick symbolic system: a system of reference and association possessing multiple levels of meaning and interpretation. I suggest that thickness, in the sense intended by symbolic anthropologists like Geertz, can be treated mathematically by bringing together two lines of formal development, that of semantic networks, and that of fractal mathematics. The resulting semantic fractals offer many advantages for modeling human culture. The properties of semantic fractals as a class are described, and their role within sociobiology and symbolic anthropology considered. Provisional empirical evidence for the hypothesis of a semantic fractal organization for culture is discussed, together with the prospects for further testing of the fractal hypothesis. Keywords: culture, culturgen, meme, fractal, semantic network. ................... MODELING THE DISTRIBUTION OF A "MEME" IN A SIMPLE AGE DISTRIBUTION POPULATION: I. A KINETICS APPROACH AND SOME ALTERNATIVE MODELS Matthew Witten Center for High Performance Computing University of Texas System, Austin, TX 78758-4497 Abstract. Although there is a growing historical body of literature relating to the mathematical modeling of social and historical processes, little effort has been placed upon modeling the spread of an idea element "meme" in such a population. In this paper we review some of the literature and we then consider a simple kinetics approach, drawn from demography, to model the distribution of a hypothetical "meme" in a population consisting of three major age groups. KEYWORDS: Meme, idea, age-structure, compartment, sociobiology, kinetics model. ................... THE PRINCIPIA CYBERNETICA PROJECT Francis Heylighen, Cliff Joslyn, and Valentin Turchin The Principia Cybernetica Project[dagger] Abstract: This note describes an effort underway by a group of researchers to build a complete and consistent system of philosophy. The system will address, issues of general philosophical concern, including epistemology, metaphysics, and ethics, or the supreme human values. The aim of the project is to move towards conceptual unification of the relatively fragmented fields of Systems and Cybernetics through consensually-based philosophical development. Keywords: cybernetics, culture, evolution, system transition, networks, hypermedia, ethics, epistemology. ................... Brain and Mind: The Ultimate Grand Challenge Elan Moritz The Institute for Memetic Research P. O. Box 16327, Panama City, Florida 32406 Abstract: Questions about the nature of brain and mind are raised. It is argued that the fundamental understanding of the functions and operation of the brain and its relationship to mind must be regarded as the Ultimate Grand Challenge problem of science. National research initiatives such as the Decade of the Brain are discussed. Keywords: brain, mind, awareness, consciousness, computers, artificial intelligence, meme, evolution, mental health, virtual reality, cyberspace, supercomputers. +=++=++=++=++=++=++=++=++=++=++=++=++=++=++=+ The Journal of Ides an archival forum for discussion of 1) evolution and spread of ideas, 2) the creative process, and 3) biological and electronic implementations of idea/knowledge generation and processing. The Journal of Ideas, ISSN 1049-6335, is published quarterly by the Institute for Memetic Research, Inc. P. O. Box 16327, Panama City Florida 32406-1327. >----------- FOR MORE INFORMATION -------> E-mail requests to Elan Moritz, Editor, at moritz at well.sf.ca.us. +=++=++=++=++=++=++=++=++=++=++=++=++=++=++=+ From jon at incsys.com Mon Apr 1 17:21:12 1991 From: jon at incsys.com (Jon Shultis) Date: Mon, 01 Apr 91 17:21:12 -0500 Subject: Informal Computing Workshop Message-ID: <9104012221.AA12203@incsys.com> Workshop on Informal Computing 29-31 May 1991 Santa Cruz, California Fundamental questions about the nature of informality are gaining importance in computer science. What is informal understanding? What is the nature of informal reasoning? Why is it so powerful and efficient? How are the inconsistency, vagueness, and incompleteness of informal thought managed? How does natural language manage to communicate informal knowledge and reasoning? Computer applications in many fields, ranging from economics and medicine to software engineering and artificial intelligence, demand effective and cognitively accurate answers to these questions in order to capture, represent, and process informal information in computer systems. Inspired by trends toward formalization in logic, mathematics, linguistics, and philosophy, computer scientists historically have tended to regard informal processes as approximate, or imperfect, realizations of formal ideals. Increasingly, however, the idea that informal languages, ontology, and reasoning can (or should) be reduced to (or supplanted by) regimented and "perfected" formalisms is being challenged. Far from being flawed formalisms, informal processes are emerging as fundamental to human understanding and language. From the "informalist" perspective, formalism has been mistaken for the paradigm of intelligence, rather than simply a useful outgrowth of intelligence. The purpose of the Workshop on Informal Computing is to define the study of Informalism, and to begin a coordinated attack on the fundamental issues and problems of the field, bringing together the insights and experience of those who have been working to understand informality in specialized domains. Discussion at the workshop will focus on three major themes: informal knowledge and reasoning; modelling and interpretation; and conversational computing and adaptive languages. Relevant topics include, but are not limited to: intentionality and consciousness; dialogue management; informal meaning and pragmatics; evidential reasoning and belief; resource- and information-limited reasoning; neurocomputation; lessons and techniques from computational linguistics; dynamical and chaotic representations and reasoning; and philosophy of language. The program will be divided between hour-long presentations by invited speakers, and discussion sessions aimed at defining and clarifying informal computing issues, and at identifying promising directions and approaches for future research. The discussion sessions should provide ample opportunity for participants to exchange views, and the schedule will be flexible enough to permit impromptu presentations as appropriate. Also, a follow-up conference may be organized if there is sufficient interest. We are busy making arrangements for speakers and drawing up the schedule, but the basic plan is to devote one day to each of the three themes mentioned above. A preliminary list of speakers includes Bruce d'Ambrosio (Oregon State University) Sandra Carberry (University of Delaware) David Fisher (Incremental Systems) Donald Good (Computational Logic) David Mundie (Incremental Systems) Larry Reeker (IDA) Jeff Rothenberg (RAND) Jon Shultis (Incremental Systems) Tim Standish (University of California at Irvine) Edward Zalta (Stanford University) The final program will be announced on or before 8 May 1991. If you are interested in participating in the workshop, please submit, by 12 April 1991, a brief summary of your interests, and previous or ongoing research that is relevant to the workshop themes. The summaries will be reviewed, and notices of acceptance sent out on 26 April 1991, together with local arrangements information. Summaries should be sent to Jon Shultis Incremental Systems Corporation 319 South Craig Street Pittsburgh, PA 15213 e-mail: jon at incsys.com tel: (412) 621-8888 FAX: (412) 621-0259 Funding for the Workshop on Informal Computing is being provided by DARPA/ISTO in conjunction with ongoing research at Incremental Systems Corporation on adaptive languages for software engineering. From RICKHAAN at ccit.arizona.edu Mon Apr 1 10:50:00 1991 From: RICKHAAN at ccit.arizona.edu (Bigamy is having one wife too many. Monogamy is the same.) Date: Mon, 1 Apr 91 08:50 MST Subject: Announcement: Conference for Philosophy and Psychology Message-ID: <9878EAB1A37F60AC4C@ccit.arizona.edu> PROGRAM, RESERVATION FORMS AND TRAVEL INFORMATION FOR SOCIETY FOR PHILOSOPHY AND PSYCHOLOGY 1991 MEETING SAN FRANCISCO STATE UNIVERSITY JUNE 9-12, 1991 MEETING SITE: Seven Hills Conference Center (west side of the San Francisco State University campus, off Lake Merced Blvd.) PROGRAM N = Nobhill Room P = Presidio Room SUNDAY, JUNE 9 9:00-11:00 Symposium I: Meaning Holism and Conceptual Role Semantics (N) Chair: Janet Levin (Philosophy, USC) Speakers: 1. Ernie Lepore and Jerry Fodor (Philosophy, Rutgers University) 2. Ned Block (Linguistics and Philosophy, MIT) 3. Michael Devitt (Philosophy, U Md/College Park) 11:00-11:15 break 11:15-12:15 Contributed paper 1 (N) Chair: Speaker: Christopher Gauker (Philosophy, University of Cincinnati) "Similarity judgments" Commentator: Robert MacCauley (Philosophy, Emory University) 11:15-12:15 Contributed paper 2 (P) Chair: Jane Duran (Philosophy, UC/Santa Barbara) Speaker: Dan Riesberg (Psychology, Reed College) and Deborah Chambers (Psychology, North Dakota State U.) "Images depict; images describe" Commentator: Kyle Cave (Psychology, Vanderbilt) 12:15-1:15 LUNCH 1:15-2:15 Contributed paper 3 Chair: Larry Birnbaum (Institute for the Learning Sciences, Northwestern University) Speaker: David Kirsh (Cognitive Science, UC/San Diego) "Some problems with the logicist view of artificial intelligence" Commentator: Pat Hayes (MCC) 1:15-2:15 Contributed paper 4 Chair: Irving Krakow (Philosophy, Camden County College) Speaker: Carol Slater (Psychology, Alma College) "Explaining behavior: why we can't lean on learning" Commentator: Joe Levine (Philosophy, NC State/Raleigh) 2:15-2:30 break 2:30-4:30 Symposium II: Is mathematics innate? (N) Chair: Speakers: 1. Karen Wynn (Psychology, U Arizona/Tucson) "Psychological evidence against empiricist theories of mathematical knowledge" 2. Robert Schwartz (Philosophy, U Wisconsin/Milwaukee) "Is mathematics innate?" 3. Philip Kitcher (Philosophy, UCSD) 4:30-4:45 break 4:45-6:00 Keynote address (N) Chair: Jerry Fodor (Philosophy, Rutgers University) Speaker: Steve Pinker (Brain and Cognitive Science, MIT) Title: "Rules and associations in human language" MONDAY, JUNE 10 9:00-11:00 Symposium III: The Roots of Social Cognition (N) Chair: Speakers: 1. Leda Cosmides (Psychology, UC/Santa Barbara) "The logic of social exchange versus the logic of logic" 2. Alan Fiske (Psychology, U Penn) "Innate hypotheses and cultural parameters for social relations" 3. Donald Symons (Anthropology, UC/Santa Barbara) 11:00-11:15 break 11:15-12:15 Contributed paper 5 Chair: Owen Flanagan (Philosophy, Wellesley College) Speaker: Janet Andrews (Psychology, Vassar College) "The role of prototypes in understanding category concepts: a critical assessment" Commentator: Margery Lucas (Psychology, Wellesley College) 11:15-12:15 Contributed paper 6 Chair: Speaker: Anthony Dardis (Philosophy, University of Georgia/Athens) "Content externalism and causal relevance" Commentator: Carol Cleland (Philosophy, University of Colorado/Boulder) 12:15-1:15 LUNCH 1:15-2:15 Contributed paper 7 Chair: Paul Pietroski (Philosophy, McGill Univ.) Speaker: Alvin Goldman (Philosophy, University of Arizona/Tucson) "Mental concepts, self-ascription, and consciousness" Commentator: Paul Boghossian (Philosophy, University of Michigan) 1:15-2:15 Contributed paper 8 Chair: Speaker: Naomi Goldblum (Psychology, Hebrew University) "The distinctiveness of metaphor: a psycholinguistic study" Commentator: Eva Kittay (Philosophy (SUNY/Stony Brook) 2:15-2:30 break 2:30-4:30 Symposium IV: Are Humans Rational? (N) Chair: ?? Speakers 1. Stuart Silvers (Philosophy, Clemson University) "Cognitive performance and assessment, epistemological norms, and the relativity of reason" 2. John Tooby (Psychology, UC/Santa Barbara) "Ecological rationality" 3. Edward Stein (Philosophy, Williams College & MIT) 7:00-8:30 Banquet (N) 8:30-10:00 Presidential Address (N) Chair: Patricia Kitcher (Philosophy, UCSD) Speaker: Ray Jackendoff (Program in Linguistics and Cognitive Science, Brandeis) Title: "Is there a faculty of social cognition?" TUESDAY, JUNE 11 9:00-10:00 Invited talk (N) Chair: Kent Bach (Philosophy, San Francisco State) Speaker: Michael Scriven (Pacific Graduate School of Psychology, Palo Alto, CA) Title: "Implicit rules in logic" 10:00-10:15 break 10:15-12:15 Symposium V: Prototype Theory and Concept Stability Chair: Ken Livingston (Psychology, Vassar College) Speakers 1. Georges Rey (Philosophy, U Md/College Park) "Concepts and reference fixers" 2. Alison Gopnik (Psychology, UC/Berkeley) "Concepts as theories: who's afraid of semantic holism" 3. George Bealer (Philosophy, UColorado/Boulder) "Philosophical constraints on a theory of concepts" 12:15-1:15 LUNCH 1:15-2:15 Contributed papers 9 (N)+(P) Chair: A.A. Howsepian (Philosophy, Notre Dame) Speakers: G. Lynn Stephens & George Graham (Philosophy, Univ of Alabama/Birmingham) "Introspective identification and disturbances of self consciousness" Commentator: Jennifer Church (Philosophy, Vassar College) 1:15-2:15 Contributed papers 10 (N)+(P) Chair: Anti Bax (Clemson University) Speaker: Kirk Ludwig (Philosophy, University of Florida/Gainesville) "How can reasons be causally relevant to what we do?" Commentator: Fred Dretske (Philosophy, Stanford University) 2:15-2:30 break 2:30-3:30 Contributed papers 11 (N)+(P) Chair: Speaker: Louise Antony (Philosophy, NC State University/Raleigh) "Rabbit pots and supernovas: on the relevance of psychological data to linguistic theory" Commentator: Scott Soames (Philosophy, Princeton University) 2:30-3:30 Contributed paper 12 (N)+(P) Chair: Speaker: Peter Ludlow (Philosophy, SUNY/Stony Brook) "In defense of the dual aspect theory" Commentator: Barbara Von Eckardt (University of Nebraska/Lincoln) 3:30-3:45 break 3:45-5:45 Contributed papers 13 & 14 (P) [B Chair: Jeffrey Poland (Psychology, U Nebraska/Lincoln) Speaker: Rob Wilson (Philosophy, Cornell University) "Individualism and causal powers" Commentator: Terry Horgan (Philosophy, Memphis State University) Speaker: Joseph Owens (Philosophy, University of Minnesota/Twin Cities) "Psychophysical supervenience and inner access" Commentator: Piers Rawling (Philosophy, University of Georgia/Athens) 3:45-5:45 Invited lecture (N) Chair: Speaker: Randy Gallistel (Psychology, UC/Berkeley) Title: "Is the associative bond the phlogiston of psychology?" Discussants: 1. Paul Churchland (Philosophy, UCSD) 2. Ben Williams (Psychology, UCSD) 5:45-6:30 Business meeting (N) WED JUNE 12 9:00-12:00 Symposium VI: Consciousness Chair: Speakers: 1. Dan Dennett (Philosophy, Tufts University) "Time and the brain: escape from the theater of consciousness" 2. Robert Van Gulick (Philosophy, Syracuse University) "Understanding the phenomenal mind" 3. Benjamin Libet (Physiology, School of Medicine, UC/San Francisco) "The cerebral time-on theory for conscious and unconscious mental function" 4. David Rosenthal (Philosophy, Graduate Center/CUNY) ........................................ Accomodations: Lodging is available at reasonable rates at two on campus locations. Both the Guest Center (hotel-like) and the Residence Halls (shared bath) are very near the Conference Center. Accomodations are available only in a complete package which includes 4 nights lodging (begining night of June 8) and breakfast and lunch each day of the conference. Rates are given on the Reservation Form (below). Limited availability for the Guest Center, so reservations will be taken in order received. (If you request the Guest Center and end up booked in the Residence Hall, the difference will be refunded at registration.) Please note that the SPP must supply SFSU with a guest list and full payment 30 days in advance to hold reservations. So please send your reservation and payment ASAP and by April 30 at the latest. Doubles can not be reserved until both persons register. STAYING ON: If you want to stay in SF for a few extra days, space should be available. Please indicate such an interest on the registration form. ARRIVAL: Check in time: after 3:00 PM Sat. June 8 (On Saturday 6/8 only - come to lounge of Verducci Hall, the only SFSU building on Lake Merced Boulevard. On other days registration will be at Conference Center) Check out time: 2:00 PM Wed. June 12 CHILDREN: Unfortunately insurance provisions do not permit children in SFSU lodging. .................................................... REGISTRATION & ROOM RESERVATIONS Room and Board (4 nights + 4 breakfasts & lunches) Guest Center: Single: $290. *Double Occ: $210. per person Residence Hall Single: $180. *Double Occ: $120. per person Total: $_____ (*If Double Occupancy, list name of person you are sharing with ___________________________) .................................................... CONFERENCE REGISTRATION: Member: $20 Nonmember: $30 Student: $10. $_________ .................................................... 1991 SPP MEMBERSHIP DUES: Regular: $15 Student: $ 5 (New members may pay 1991 dues and register as members.) $_________ ................................................... PRESIDENTIAL ADDRESS BANQUET June 10 Price: $18.00 $_________ .................................................... PARKING: $6.00 $_______ _ GRAND TOTAL $_________ Your Name______________________________ Address________________________________ MAKE CHECK PAYABLE TO SOCIETY FOR PHILOSOPHY AND PSYCHOLOGY (WILL NOT BE CASHED BEFORE MAY 1) AND MAIL ASAP (AND NO LATER THAN APRIL 30) TO: KENT BACH 66 ST. GERMAIN SAN FRANCISCO, CA 94114 Check here if you wish to know about the possibilities of staying on for extra days and indicate dates _________ GETTING TO SAN FRANCISCO STATE UNIVERSITY BY AIR: Fly to San Franciso International Airport. Shuttles are readily available outside the arrivals area. Look for vans marked SF and just ask the driver if his/her van goes to San Francisco State University. BY CAR: From the North: Take Highway 1 south to 19th and Holloway Avenues. From the South: Take Interstate 280 north; exit at 19th Ave. Take Junipero Serra to Holloway Ave; turn left onto Holloway. From the East: Take Interstate 80 across the Oakland Bay Bridge to Route 101 south to Interstate 280. Take 19th Ave. exit; bear right onto Sagamore Blvd. to Brotherhood Way to Junipero Serra Blvd north. Take Junipero Serra to Holloway Ave., turn left on Holloway to 19th Ave. LOCAL TRANSIT: From downtown San Francisco, take the Muni No. 17 Express (2nd and Market Streets) or the M Streetcar. For further information about any aspect of local arrangements call Kent Bach at 415-665-1040 The Program Chairs wish to thank the following for their help in refereeing contributed papers: Louise Antony, Kent Bach, Ned Block, Denise Cummins, A.J. Figuerado, Owen Flanagan, Alvin Goldman, George Graham, Kerry Green, Eli Hirsch, Terry Horgan, Eva Kittay, Ernie LePore, Dan Lloyd, Chris Maloney, Charles Marks, Bob Matthews, Lynn Nadel, Janet Nicol, Sarah Patterson, Paul Pietroski, William Rappoport, Georges Rey, Mark Richard, Bob Schwartz, Stuart Silvers, Varda Solomon, Ed Stein, Bob Van Gulik, Barbara Von Eckhardt, Stephen White, David Wong, Karen Wynn, Palle Yourgrau Program Chairs: Jerry Samet (Philosophy, Brandeis University) Paul Bloom (Psychology, University of Arizona) E-mail BLOOM at ccit.arizona.edu From ps_coltheart at vaxa.mqcc.mq.oz.au Wed Apr 3 08:44:23 1991 From: ps_coltheart at vaxa.mqcc.mq.oz.au (Max Coltheart) Date: Wed, 3 Apr 91 08:44:23 est Subject: Full Professor Position in Australia Message-ID: <9104022244.AA00985@macuni.mqcc.mq.oz.au> PROFESSOR OF PSYCHOLOGY VACANCY Applications are invited for appointment to the position of Professor of Psychology at Macquarie University in Sydney, NSW, Australia. Applicants with interests in any area of psychology will be considered. Professor M. Coltheart and Professor J. K. Collins currently hold chairs within the department. Further information about the University, conditions of appointment and method of application may be obtained from the Academic Staff Office, Macquarie University, NSW 2109 or by telephoning (02) 805-7390. Closing date for applications 14 June 1991. Initial enquiries may be directed to Professor Max Coltheart (ps_coltheart at vaxa.mqcc.mq.oz.au, telephone (02) 805-8086, confidential fax (02) 805-8127) or Professor John Collins (tel (02) 805-8030) or Associate Professor George Cooney (ps_cooney at vaxa.mqcc.mq.oz.au, telephone (02) 805-8067). From pratt at paul.rutgers.edu Fri Apr 5 17:08:52 1991 From: pratt at paul.rutgers.edu (Lorien Y. Pratt) Date: Fri, 5 Apr 91 17:08:52 EST Subject: Two papers on information transfer / problem decomposition Message-ID: <9104052208.AA08551@paul.rutgers.edu> The following two papers are now available via FTP from the neuroprose archives. The first is for AAAI91, so written towards an AI/Machine learning audience. The second is for IJCNN91, so more neural network-oriented. There is some overlap between them: the AAAI paper reports briefly on the study describved in more detail in the IJCNN paper. Instructions for retrieval are at the end of this message. --Lori #@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@#@ Direct Transfer of Learned Information Among Neural Networks To appear: Proceedings of AAAI-91 Lorien Y. Pratt and Jack Mostow and Candace A. Kamm Abstract A touted advantage of symbolic representations is the ease of transferring learned information from one intelligent agent to another. This paper investigates an analogous problem: how to use information from one neural network to help a second network learn a related task. Rather than translate such information into symbolic form (in which it may not be readily expressible), we investigate the direct transfer of information encoded as weights. Here, we focus on how transfer can be used to address the important problem of improving neural network learning speed. First we present an exploratory study of the somewhat surprising effects of pre-setting network weights on subsequent learning. Guided by hypotheses from this study, we sped up back-propagation learning for two speech recognition tasks. By transferring weights from smaller networks trained on subtasks, we achieved speedups of up to an order of magnitude compared with training starting with random weights, even taking into account the time to train the smaller networks. We include results on how transfer scales to a large phoneme recognition problem. @%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@%@ Improving a Phoneme Classification Neural Network through Problem Decomposition To appear: Proceedings of IJCNN-91 L. Y. Pratt and C. A. Kamm Abstract In the study of neural networks, it is important to determine whether techniques that have been validated on smaller experimental tasks can be scaled to larger real-world problems. In this paper we discuss how a methodology called {\em problem decomposition} can be applied to AP-net, a neural network for mapping acoustic spectra to phoneme classes. The network's task is to recognize phonemes from a large corpus of multiple-speaker, continuously-spoken sentences. We review previous AP-net systems and present results from a decomposition study in which smaller networks trained to recognize subsets of phonemes are combined into a larger network for the full signal-to-phoneme mapping task. We show that, by using this problem decomposition methodology, comparable performance can be obtained in significantly fewer arithmetic operations. ^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^%^% To retrieve: unix> ftp cheops.cis.ohio-state.edu (or 128.146.8.62) Name: anonymous Password: neuron ftp> cd pub/neuroprose ftp> binary ftp> get pratt.aaai91.ps.Z ftp> get pratt.ijcnn91.ps.Z ftp> quit unix> uncompress pratt.aaai91.ps.Z pratt.ijcnn91.ps.Z unix> lpr pratt.aaai91.ps pratt.ijcnn91.ps From ernst at acquine.cns.caltech.edu Fri Apr 5 23:11:43 1991 From: ernst at acquine.cns.caltech.edu (Ernst Niebur) Date: Fri, 5 Apr 91 20:11:43 PST Subject: Connectionist Faculty Position at Boulder In-Reply-To: Paul Smolensky's message of Tue, 23 Oct 90 10:04:58 -0600 <9010231604.AA06346@axon.Colorado.EDU> Message-ID: <9104060411.AA02276@cns.caltech.edu> Please note my new internet address (ernst at cns.caltech.edu instead of the former ernst at descartes.cns.caltech.edu). Also, in the course of changing our system, some messages were lost. If you sent me a message between March 18 and April 5 which I did not answer, it means the message was lost. Please send it again. Sorry for the inconvenience. --Ernst Niebur ========================================================================= Ernst Niebur Phone (818)356-6885 Computation and Neural Systems Fax (818)796-8876 Caltech 216-76 ernst at cns.caltech.edu Pasadena, CA 91125 ernst at caltech.bitnet USA ========================================================================= From ibm at dit.upm.es Mon Apr 8 07:08:01 1991 From: ibm at dit.upm.es (Ignacio Bellido Montes) Date: Mon, 8 Apr 91 13:08:01 +0200 Subject: Neural Nets Workshop (IWANN 91) Message-ID: <9104081108.AA01055@bosco.dit.upm.es> There will be a Neural Nets Workshop Granada (Spain) next September. I think this is the first international workshop of this kind that will be held in Spain, I hope all of you can send contributions and participate. Here I send you the electronic version of the CFP followed by the LaTeX version. If you have any question about the workshop, please ask me or the people of the workshop secretary. I'm not on the organization but I can help finding people etc... About Granada, I can tell you this is one of the most wonderful cities in the Spain and ... in the World. It has one of the most famous and wonder arab contributions to the culture, "La Alhambra", and small and quiet lanes on the old side of the city, the "Albaicin", that transmit the visitor a very good feeling. I hope you be able to participate and enjoy... See you in Granada. Gregorio Fernandez Dpt. Ingenieria de Sistemas Telematicos ETSI Telecomunicacion- UPM Ciudad Universitaria 28040 Madrid, Spain E-mail: gfernandez at dit.upm.es ------------------------------------------------------------------------------- INTERNATIONAL WORKSHOP ON ARTIFICIAL NEURAL NETWORKS IWANN'91 First Announcement and Call for Papers Granada, Spain September 17-19, 1991 ORGANISED AND SPONSORED BY Spanish Chapter of the Computer Society of the IEEE, AEIA (IEEE Affiliate Society), and Department of Electronic and Computer Technology. University of Granada. Spain. SCOPE Artificial Neural Networks (ANN) were first developed as structural or functional modelling systems of natural ones, featuring the ability to perform problem-solving tasks. They can be thought as computing arrays consisting of series of repetitive uniform processors (neuron-like elements) placed on a grid. Learning is achieved by changing the interconnections between these processing elements. Hence, these systems are also called connectionist models. ANN has become a subject of wide-spread interest: they offer an odd scheme-based programming standpoint and exhibit higher computing speeds than conventional von-Neumann architectures, thus easing or even enabling handling complex task such as artificial vision, speech recognition, information recovery in noisy environments or general pattern recognition. In ANN systems, collective information management is achieved by means of parallel operation of neuron-like elements, into which information processing is distributed. It is intended to exploit this highly parallel processing capability as far as possible in complex problem-solving tasks. Cross-fertilization between the domains of artificial and real neural nets is desirable. The more genuine problems of biological computation and information processing in the nervous system still remain open and contributions in this line are more than welcome. Methodology, theoretical frames, structural and organizational principles in neuroscience, self-organizing and co- operative processes and knowledge based descriptions of neural tissue are relevant topics to bridge the gap between the artificial and natural perspectives. The workshop intends to serve as a meeting place for engineers and scientists working in this area, so that present contacts and relationships can be further increased. The workshop will comprise two complementary activities: . scientific and technical conferences, and . scientific communications sessions. TOPICS The workshop is open to all aspects of artificial neural networks, including: 1. Neural network theories. Neural models. 2. Biological perspectives 3. Neural network architectures and algorithms. 4. Software developments and tools. 5. Hardware implementations 6. Applications. LOCATION Facultad de Ciencias Campus Universitario de Fuentenueva Universidad de Granada 18071 GRANADA. (SPAIN) LANGUAGES English and Spanish will be the official working languages. English is preferable as the working language. CALL FOR PAPERS The Programme Committee seeks original papers on the six above mentioned areas. Survey papers on the various available approaches or particular application domains are also sought. In their submitted papers, authors should pay particular attention to explaining the theoretical and technical choices involved, to make clear the limitations encountered and to describe the current state of development of their work. INSTRUCTIONS TO AUTHORS Three copies of submitted papers (not exceeding 8 pages in 21x29.7 cms (DIN-A4), with 1,6 cm. left, right, top and bottom margins) should be received by the Programme Chairman at the address below before June 20, 1991. The headlines should be centered and include: . the title of paper in capitals . the name(s) of author(s) . the address(es) of author(s), and . a 10 line abstract. Three blank lines should be left between each of the above items, and four between the headlines and the body of the paper, written in English, single-spaced and not exceeding the 8 pages limit. All papers received will be refereed by the Programme Committee. The Committee will communicate their decision to the authors on July 10. Accepted papers will be published in the proceedings to be distributed to workshop participants. In addition to the paper, one sheet should be attached including the following information: . the title of the paper, . the name(s) of author(s), . a list of five keywords, . a reference to which of the six topics the paper concerns, and . postal address of one of the authors, with phone and fax numbers, and E-mail (if available). We intend to get in touch with various international publishers (such as Springer-Verlag and Prentice-Hall) for the final version of the proceedings. Contributions to be sent to: Prof. Jose Mira Dpto. Informatica y Automatica UNED C/Senda del Rey s/n 28040 MADRID (Spain) Tel. (34) 1 5 44 60 00 Fax: (34) 1 5 44 67 37 ACCOMMODATION A list of available Hotels will be sent on registration. Hotel reservations can be made directly by each participant with the local agency below. All request should be addressed to: Viajes Internacional Expreso (VIE) Galerias Preciados Carrera del Genil, s/n 18005 GRANADA (Spain) Tel. (34) 58-22.44.95, -22.75.86, -224944 Telex: 78525 We can only guarantee to accept reservation received by July 25. REGISTRATION FEE . Regular fee: 35.000 ptas. . IEEE, AEIA and ATI members fee: 30.000 ptas. . Scholarship holders fee: 5.000 ptas. Inscription payments: Transfer to: IWANN'91 account number: 16.142.512 Caja Postal (Code: 2088-2037.1) Camino de Ronda, 138 18003 GRANADA (SPAIN) or alternatively, cheque made out to: IWANN'91 (16.142.512) Secretariat address: Departamento de Electronica y Tecnologia de Computadores Facultad de Ciencias Universidad de Granada 18071 GRANADA (SPAIN) FAX: 34-58-24.32.30 or 34-58-27.42.58 Phone: 34-58-24.32.26 E-Mail: jmerelo at ugr.es aprieto at ugr.es PROGRAM AND ORGANIZATION COMMITTEE Senen Barro Univ. de Santiago Joan Cabestany Univ. Pltca. de Catalunya Jose Antonio Corrales Univ. Oviedo. Gregorio Fernandez Univ. Pltca. de Madrid J. Simoes da Fonseca Univ. de Lisboa Antonio Lloris Univ. Granada Javier Lopez Aligue Univ. de Extremadura. Jose Mira (Programme Chairman) UNED. Madrid Roberto Moreno Univ Las Palmas Gran Canaria Alberto Prieto (Organization Chairman) Univ. Granada Francisco Sandoval Univ. de Malaga Carmen Torras Instituto de Cibernetica. CSIC. Barcelona Elena Valderrama CNM- Univ. Autonoma de Barcelona LOCAL ORGANIZING COMMITTEE (Universidad de Granada) Juan Julian Merelo Julio Ortega Francisco J. Pelayo Begona del Pino Alberto Prieto (To be completed and returned as soon as possible to: Departamento de Electronica. Facultad de Ciencias. Univ. de Granada. 18071 GRANADA (SPAIN). FAX (34)-58-24.32.30) Block letters, please Name: Company/Organization: Address: State/Country: E-mail: Phone: Fax: Please tick as appropriate: I intend to: attend to the workshop submit a paper Name(s) of Author(s): Provisional Title: ------------------------------------------------------------------------------ %%Latex-version \documentstyle{article} \setlength{\textheight}{24.0cm} \setlength{\textwidth}{16.0cm} \setlength{\voffset}{-1.0in} \setlength{\hoffset}{-1.0in} \begin{document} \thispagestyle{empty} \vspace*{3cm} \begin{center} {\Large \bf INTERNATIONAL WORKSHOP\\ ON\\ ARTIFICIAL NEURAL NETWORKS\\ \vspace{2cm} IWANN'91\\} {\large \vspace{4cm} First Announcement and Call for Papers\\ \vspace{4cm} Granada, Spain\\ September 17-19, 1991\\ \vspace{4cm} ORGANISED AND SPONSORED BY:\\ Spanish Chapter of the Computer Society of the IEEE,\\ AEIA (IEEE Affiliate Society), and\\ Department of Electronic and Computer Technology\\ (University of Granada, Spain)} \end{center} \newpage \section*{SCOPE} Artificial Neural Networks (ANN) were first developed as structural or functional modelling systems of natural ones, featuring the ability to perform problem-solving tasks. They can be thought as computing arrays consisting of series of repetitive uniform processors (neuron-like elements) placed on a grid. Learning is achieved by changing the interconnections between these processing elements. Hence, these systems are also called connectionist models. ANN has become a subject of wide-spread interest: they offer an odd scheme-based programming standpoint and exhibit higher computing speeds than conventional von-Neumann architectures, thus easing or even enabling handling complex task such as artificial vision, speech recognition, information recovery in noisy environments or general pattern recognition. In ANN systems, collective information management is achieved by means of parallel operation of neuron-like elements, into which information processing is distributed. It is intended to exploit this highly parallel processing capability as far as possible in complex problem-solving tasks. Cross-fertilization between the domains of artificial and real neural nets is desirable. The more genuine problems of biological computation and information processing in the nervous system still remain open and contributions in this line are more than welcome. Methodology, theoretical frames, structural and organizational principles in neuroscience, self-organizing and co- operative processes and knowledge based descriptions of neural tissue are relevant topics to bridge the gap between the artificial and natural perspectives. The workshop intends to serve as a meeting place for engineers and scientists working in this area, so that present contacts and relationships can be further increased. The workshop will comprise two complementary activities: \begin{itemize} \item scientific and technical conferences, and \item scientific communications sessions. \end{itemize} \section*{TOPICS} The workshop is open to all aspects of artificial neural networks, including: \begin{enumerate} \item Neural network theories. Neural models. \item Biological perspectives \item Neural network architectures and algorithms. \item Software developments and tools. \item Hardware implementations \item Applications. \end{enumerate} \section*{LOCATION} Facultad de Ciencias\\ Campus Universitario de Fuentenueva\\ Universidad de Granada\\ 18071 GRANADA (SPAIN)\\ \section*{LANGUAGES} English and Spanish will be the official working languages. English is preferable as the working language. \newpage \section*{CALL FOR PAPERS} The Programme Committee seeks original papers on the six above mentioned areas. Survey papers on the various available approaches or particular application domains are also sought. In their submitted papers, authors should pay particular attention to explaining the theoretical and technical choices involved, to make clear the limitations encountered and to describe the current state of development of their work. \section*{INSTRUCTIONS TO AUTHORS} Three copies of submitted papers (not exceeding 8 pages in 21x29.7 cms (DIN-A4), with 1,6 cm. left, right, top and bottom margins) should be received by the Programme Chairman at the address below before June 20, 1991. The headlines should be centered and include:\\ $\bullet$ the title of paper in capitals\\ $\bullet$ the name(s) of author(s)\\ $\bullet$ the address(es) of author(s), and\\ $\bullet$ a 10 line abstract. Three blank lines should be left between each of the above items, and four between the headlines and the body of the paper, written in English, single-spaced and not exceeding the 8 pages limit. All papers received will be refereed by the Programme Committee. The Committee will communicate their decision to the authors on July 10. Accepted papers will be published in the proceedings to be distributed to workshop participants. In addition to the paper, one sheet should be attached including the following information:\\ $\bullet$ the title of the paper,\\ $\bullet$ the name(s) of author(s),\\ $\bullet$ a list of five keywords,\\ $\bullet$ a reference to which of the six topics the paper concerns, and\\ $\bullet$ postal address of one of the authors, with phone and fax numbers, and E-mail (if available). \vspace{0.5cm} We intend to get in touch with various international publishers (such as Springer-Verlag and Prentice-Hall) for the final version of the proceedings. \vspace{0.5cm} \noindent {\bf Contributions to be sent to:} \noindent Prof. Jos\'{e} Mira\\ UNED---Dpto. Inform\'{a}tica y Autom\'{a}tica\\ C/Senda del Rey s/n\\ 28040 MADRID (Spain)\\ Tel. (34) 1 5 44 60 00; Fax: (34) 1 5 44 67 37\\ \section*{ACCOMMODATION} A list of available Hotels will be sent on registration. Hotel reservations can be made directly by each participant with the local agency below. All request should be addressed to:\\ \noindent Viajes Internacional Expreso (VIE)\\ Galerias Preciados\\ Carrera del Genil, s/n\\ 18005 GRANADA (Spain)\\ Tel. (34) 58-22.44.95, -22.75.86, -224944\\ Telex: 78525\\ We can only guarantee to accept reservation received by July 25. \newpage \section*{REGISTRATION FEE} \begin{tabular}{lr} Regular fee: & 35.000 ptas.\\ IEEE, AEIA and ATI members fee: & 30.000 ptas.\\ Scholarship holders fee: & 5.000 ptas.\\ \end{tabular} \subsection*{Inscription payments: } \begin{tabular}{ll} Transfer to: & IWANN'91\\ & account number: 16.142.512\\ & Caja Postal (Code: 2088-2037.1)\\ & Camino de Ronda, 138\\ & 18003 GRANADA (SPAIN)\\ or alternatively, cheque made out to: & IWANN'91 (16.142.512)\\ \end{tabular} \subsection*{Secretariat address:} Departamento de Electr\'{o}nica y Tecnolog\'{i}a de Computadores\\ Facultad de Ciencias\\ Universidad de Granada\\ 18071 GRANADA (SPAIN)\\ FAX: 34-58-24.32.30 or 34-58-27.42.58\\ Phone: 34-58-24.32.26\\ E-Mail: jmerelo at ugr.es, or: aprieto at ugr.es \section*{PROGRAM AND ORGANIZATION COMMITTEE} \begin{tabular}{lr} Sen\'{e}n Barro & Univ. de Santiago\\ Joan Cabestany & Univ. Pltca. de Catalunya\\ Jose Antonio Corrales& Univ. Oviedo\\ Gregorio Fern\'{a}ndez & Univ. Pltca. de Madrid\\ J. Simoes da Fonseca & Univ. de Lisboa\\ Antonio Lloris & Univ. Granada\\ Javier L\'{o}pez Aligu\'{e} & Univ. de Extremadura.\\ Jos\'{e} Mira {\bf (Programme Chairman)} & UNED. Madrid\\ Roberto Moreno & Univ. Las Palmas de Gran Canaria\\ Alberto Prieto {\bf (Organization Chairman)}& Univ. Granada\\ Francisco Sandoval & Univ. de M\'{a}laga\\ Carme Torras & Instituto de Cibern\'{e}tica. CSIC. Barcelona\\ Elena Valderrama & CNM--Univ. Aut\'{o}noma de Barcelona\\ \end{tabular} \section*{LOCAL ORGANIZING COMMITTEE} {\large \bf(Universidad de Granada)}\\ Juan Juli\'{a}n Merelo\\ Julio Ortega\\ Francisco J. Pelayo\\ Bego\~{n}a del Pino\\ Alberto Prieto \\ \newpage \noindent (To be completed and returned as soon as possible to:\\ Departamento de Electronica, Facultad de Ciencias,\\ Univ. de Granada, 18071 GRANADA, SPAIN,\\ FAX (34)-58-24.32.30)\\ Block letters, please\\ \vspace{0.5cm} Name:\\ Company/Organization:\\ Address:\\ \vspace{1.0cm} State/Country:\\ E-mail: \hspace{4cm} Phone: \hspace{4cm} Fax:\\ Please tick as appropriate:\\ I intend to:\\ $\Box$ attend to the workshop\\ $\Box$ submit a paper\\ Name(s) of Author(s):\\ Provisional Title: \end{document} From SAYEGH at CVAX.IPFW.INDIANA.EDU Wed Apr 10 20:41:22 1991 From: SAYEGH at CVAX.IPFW.INDIANA.EDU (SAYEGH@CVAX.IPFW.INDIANA.EDU) Date: Wed, 10 Apr 1991 20:41:22 EDT Subject: Proceedings of Third NN and PDP Conference, Indiana-Purdue Message-ID: <910410204122.20204a59@CVAX.IPFW.INDIANA.EDU> The proceedings of the THIRD conference on Neural Networks and Parallel Distributed Processing held in April 1990 at Indiana-Purdue University in Ft Wayne can be obtained by writing to: Ms. Sandra Fisher Physics Department Indiana University-Purdue University Ft Wayne, IN 46805 and including $5 + $1 for mailing and handling. Checks should be made payable to The Indiana-Purdue Foundation. The 109 page proceedings contain the following articles: INTEGRATED AUTONOMOUS NAVIGATION BY ADAPTIVE NEURAL NETWORKS Dean A. Pomerleau Department of Computer Science Carnegie Mellon University APPLYING A HOPFIELD-STYLE NETWORK TO DEGRADED PRINTED TEXT RESTORATION Arun Jagota Department of Computer Science State University of New York at Buffalo RECENT STUDIES WITH PARALLEL, SELF-ORGANIZ- ING, HIERARCHICAL NEURAL NETWORKS O.K. Ersoy & D. Hong School of Electrical Engineering Purdue University INEQUALITIES, PERCEPTRONS AND ROBOTIC PATH- PLANNING Samir I. Sayegh Department of Physics Indiana University-Purdue University GENETIC ALGORITHMS FOR FEATURE SELECTION FOR COUNTERPROPAGATION NETWORKS F.Z. Brill & W.N. Martin Department of Computer Science University of Virginia MULTI-SCALE VISION-BASED NAVIGATION ON DIS- TRIBUTED-MEMORY MIMD COMPUTERS A.W. Ho & G.C. Fox Caltech Concurrent Computation Program California Institute of Technology A NEURAL NETWORK WHICH ENABLES SPECIFICATION OF PRODUCTION RULES N. Liu & K.J. Cios The University of Toledo PIECE-WISE LINEAR ESTIMATION OF MECHANICAL PROPERTIES OF MATERIALS WITH NEURAL NETWORKS I.H. Shin, K.J. Cios, A. Vary* & H.E. Kautz* The University of Toledo & NASA Lewis Re- search Center* INFLUENCE OF THE COLUMN STRUCTURE ON INTRA- CORTICAL LONG RANGE INTERACTIONS E. Niebur & F. Worgotter California Institute of Technology LEARNING BY GRADIENT DESCENT IN FUNCTION SPACE Ganesh Mani University of Wisconsin-Madison REAL TIME DYNAMIC RECOGNITION OF SPATIAL TEMPORAL PATTERNS M. F. Tenorio School of Electrical Engineering Purdue University A NEURAL ARCHITECTURE FOR COGNITIVE MAPS Martin Sonntag Cognitive Science & Machine Intelligence Lab University of Michigan P.S. The Fourth Conference is scheduled to start April 11, 91 at 6pm in the Classroom Medical Building, CM159, of the Fort Wayne Campus of Indiana and Purdue. A previous announcement of this conference was made on the list. From andreas%psych at Forsythe.Stanford.EDU Thu Apr 11 00:48:16 1991 From: andreas%psych at Forsythe.Stanford.EDU (Andreas Weigend) Date: Wed, 10 Apr 91 21:48:16 PDT Subject: No subject Message-ID: <9104110448.AA07648@psych> ~t Subject: prepint by Lumer & Huberman: "Binding Hierarchies:..." The following preprint is available in hardcopy form. It can be obtained by sending e-mail to: lumer at parc.xerox.com Please do NOT reply to this message. ______________________________________________________________________ Binding Hierarchies: A Basis for Dynamic Perceptual Grouping E. Lumer and B. A. Huberman Stanford University and Xerox PARC Abstract Since it has been suggested that the brain binds its fragmentary representations of perceptual events via phase-locking of stimulated neuron oscillators, it is important to determine how extended synchronization can occur in a clustered organization of cells possessing a distribution of firing rates. In order to answer that question, we establish the basic conditions for the existence of a binding mechanism based on phase-locked oscillations. In addition, we present a simple hierarchical architecture of feedback units which not only induces robust synchronization within and segregation between perceptual groups, but also serves as a generic binding machine. ______________________________________________________________________ From johni at ee.ubc.ca Wed Apr 10 15:25:53 1991 From: johni at ee.ubc.ca (johni@ee.ubc.ca) Date: Wed, 10 Apr 91 12:25:53 PDT Subject: NIPS 91 information request Message-ID: <9104101925.AA22205@snickers.ee.ubc.ca> Hi, does anybody have information on NIPS '91? I like to know the time, place, contact, dealine for submission of papers, ... Thanks. From leow at mcc.com Thu Apr 11 12:37:53 1991 From: leow at mcc.com (Wee Kheng Leow) Date: Thu, 11 Apr 91 11:37:53 CDT Subject: sharing hotel room Message-ID: <9104111637.AA23020@koala.aca.mcc.com> I'm attending the course and conference on neural networks (Neural Networks: From Foundations to Applications, and Neural Networks for Vision and Image Processing) to be held in Tyngsboro, MA on May 5-12. I've reserved a double room in Red Roof Inn (Nashua, NH) and like to look for a partner to share the room. Interested participant please email to leow at mcc.com (DO NOT reply to this message). Wee Kheng Leow p.s. I'm a male Chinese graduate student and I speak English and Mandarin. From collins at z.ils.nwu.edu Thu Apr 11 12:28:13 1991 From: collins at z.ils.nwu.edu (Gregg Collins) Date: Thu, 11 Apr 91 11:28:13 CDT Subject: Conference announcement: ML91 Message-ID: <9104111628.AA03045@z.ils.nwu.edu> ML91 -- The Eighth International Workshop on Machine Learning Registration Information On behalf of the organizing committee, and the individual workshop committees, we are pleased to announce that ML91, the Eighth International Workshop on Machine Learning, will be held at Northwestern University, Evanston, Illinois, USA, June 27-29, 1991. ML91 comprises eight workshop tracks: o Automated Knowledge Acquisition o Computational Models of Human Learning o Constructive Induction o Learning from Theory and Data o Learning in Intelligent Information Retrieval o Learning Reaction Strategies o Learning Relations o Machine Learning in Engineering Automation In addition there will be plenary talks by noted researchers in machine learning and related fields. Registration The registration fee will be $70 for students, $100 for everyone else. This fee will cover conference participation, proceedings, and receptions. To register please send a check for the appropriate amount, made out to Northwestern University, along with a completed copy of the form to be found at the end of this message. The deadline for registration is May 22, 1991. After this date, a late fee of $25 will be charged. Accomodations We have reserved rooms at the following hotels: Omni Orrington Hotel 1710 Orrington Avenue Evanston, Illinois 60201 (708) 866-8700 or 1-800-THE-OMNI Rates: $78.00 single per night $88.00 double per night The Orrington is a five minute walk from the conference site. Holiday Inn -- Evanston 1501 Sherman Avenue Evanston, Illinois 60201 (708) 491-6400 or 1-800-HOLIDAY Rates: $60.00 single per night $70.00 double per night The Holiday Inn is a fifteen minute walk from the conference site. You must make hotel reservations yourself. When you reserve your room, please inform the hotel that you are registering for the Machine Learning Workshop. We are currently trying to arrange dorm space on campus. A message about this will be forthcoming shortly. Travel By air: Chicago has two airports: O'Hare and Midway. Most flights go to O'Hare, which is also the most convenient to Evanston. To get from O'Hare to Evanston, the following options are available: Taxi: The fare from O'Hare to Evanston should be about $25. Bus Service: Continental Air Tranport (312-454-7799) and Larry Webb Bus Service (312-866-7163) leave O'Hare for Evanston every hour from the American Airlines baggage area. Each line will take you to the Omni Orrington Hotel in Evanston. The Holiday Inn is 3 blocks south of the Orrington. Public Transportation: The CTA (1-800-972-7000) "El" trains run from O'Hare to downtown Chicago, and from Chicago to Evanston. Follow the airport signs to locate the O'Hare "El" stop. The fare is $1 per person. Take the "El" to the Washington St. stop, which is the end of the line. Transfer to a Northbound train. Take this train to the Howard St. Station. Transfer to a Northbound Evanston train. Get off at the Davis St. station in Evanston. To get to the Holiday Inn, walk east to Sherman Ave., turn right and go south two blocks. To get to the Orrington, walk east to Orrington Ave., turn left and go north one block. This trip will take at least an hour. It's reasonably safe but we don't recommend it because of the aggravation involved. By car: To get to Evanston, take Dempster St. east from either the Tri-State Tollway (I-294) or the Edens Expressway (I-94). Follow Dempster east to Chicago Ave. (one block past the "El" train overpass). Turn left on Chicago and go north. To get to the Holiday Inn, turn left on Grove St., go one block west. To get to the Orrington, turn left on Davis St., go one block west to Orrington, turn right and go one and a half blocks north to the hotel. By train: Amtrack trains stop at Union Station in downtown Chicago. Walk 4 blocks north on Canal St. to the Chicago & Northwestern Commuter Train Station. Take a Northbound train to the Davis St. station in Evanston. Directions from there to the hotels are as from the "El" above. **************************Registration Form************************** Please send this form, along with a check in the appropriate amount made out to Northwestern University, to the following address: Machine Learning 1991 The Institute for the Learning Sciences 1890 Maple Avenue Evanston, Illinois, 60201 USA phone (708) 491-3500 fax (708) 491-5258 email ml91 at ils.nwu.edu Registration information (please type or print): Name: Address: Phone: Email: Type of registration: ( ) Student -- $70 ( ) Others -- $100 Registration is due May 22, 1991. If your registration will arrive after that date, please add a late fee of $25. From collins%z.ils.nwu.edu at VM.TCS.Tulane.EDU Thu Apr 11 12:28:13 1991 From: collins%z.ils.nwu.edu at VM.TCS.Tulane.EDU (Gregg Collins) Date: Thu, 11 Apr 91 11:28:13 CDT Subject: Conference announcement: ML91 Message-ID: <9104111628.AA03045@z.ils.nwu.edu> ML91 -- The Eighth International Workshop on Machine Learning Registration Information On behalf of the organizing committee, and the individual workshop committees, we are pleased to announce that ML91, the Eighth International Workshop on Machine Learning, will be held at Northwestern University, Evanston, Illinois, USA, June 27-29, 1991. ML91 comprises eight workshop tracks: o Automated Knowledge Acquisition o Computational Models of Human Learning o Constructive Induction o Learning from Theory and Data o Learning in Intelligent Information Retrieval o Learning Reaction Strategies o Learning Relations o Machine Learning in Engineering Automation In addition there will be plenary talks by noted researchers in machine learning and related fields. Registration The registration fee will be $70 for students, $100 for everyone else. This fee will cover conference participation, proceedings, and receptions. To register please send a check for the appropriate amount, made out to Northwestern University, along with a completed copy of the form to be found at the end of this message. The deadline for registration is May 22, 1991. After this date, a late fee of $25 will be charged. Accomodations We have reserved rooms at the following hotels: Omni Orrington Hotel 1710 Orrington Avenue Evanston, Illinois 60201 (708) 866-8700 or 1-800-THE-OMNI Rates: $78.00 single per night $88.00 double per night The Orrington is a five minute walk from the conference site. Holiday Inn -- Evanston 1501 Sherman Avenue Evanston, Illinois 60201 (708) 491-6400 or 1-800-HOLIDAY Rates: $60.00 single per night $70.00 double per night The Holiday Inn is a fifteen minute walk from the conference site. You must make hotel reservations yourself. When you reserve your room, please inform the hotel that you are registering for the Machine Learning Workshop. We are currently trying to arrange dorm space on campus. A message about this will be forthcoming shortly. Travel By air: Chicago has two airports: O'Hare and Midway. Most flights go to O'Hare, which is also the most convenient to Evanston. To get from O'Hare to Evanston, the following options are available: Taxi: The fare from O'Hare to Evanston should be about $25. Bus Service: Continental Air Tranport (312-454-7799) and Larry Webb Bus Service (312-866-7163) leave O'Hare for Evanston every hour from the American Airlines baggage area. Each line will take you to the Omni Orrington Hotel in Evanston. The Holiday Inn is 3 blocks south of the Orrington. Public Transportation: The CTA (1-800-972-7000) "El" trains run from O'Hare to downtown Chicago, and from Chicago to Evanston. Follow the airport signs to locate the O'Hare "El" stop. The fare is $1 per person. Take the "El" to the Washington St. stop, which is the end of the line. Transfer to a Northbound train. Take this train to the Howard St. Station. Transfer to a Northbound Evanston train. Get off at the Davis St. station in Evanston. To get to the Holiday Inn, walk east to Sherman Ave., turn right and go south two blocks. To get to the Orrington, walk east to Orrington Ave., turn left and go north one block. This trip will take at least an hour. It's reasonably safe but we don't recommend it because of the aggravation involved. By car: To get to Evanston, take Dempster St. east from either the Tri-State Tollway (I-294) or the Edens Expressway (I-94). Follow Dempster east to Chicago Ave. (one block past the "El" train overpass). Turn left on Chicago and go north. To get to the Holiday Inn, turn left on Grove St., go one block west. To get to the Orrington, turn left on Davis St., go one block west to Orrington, turn right and go one and a half blocks north to the hotel. By train: Amtrack trains stop at Union Station in downtown Chicago. Walk 4 blocks north on Canal St. to the Chicago & Northwestern Commuter Train Station. Take a Northbound train to the Davis St. station in Evanston. Directions from there to the hotels are as from the "El" above. **************************Registration Form************************** Please send this form, along with a check in the appropriate amount made out to Northwestern University, to the following address: Machine Learning 1991 The Institute for the Learning Sciences 1890 Maple Avenue Evanston, Illinois, 60201 USA phone (708) 491-3500 fax (708) 491-5258 email ml91 at ils.nwu.edu Registration information (please type or print): Name: Address: Phone: Email: Type of registration: ( ) Student -- $70 ( ) Others -- $100 Registration is due May 22, 1991. If your registration will arrive after that date, please add a late fee of $25. From cuny91 at prodigal.psych.rochester.edu Thu Apr 11 15:27:43 1991 From: cuny91 at prodigal.psych.rochester.edu (A. Zukowski (CUNY Coordinator)) Date: Thu, 11 Apr 91 15:27:43 EDT Subject: No subject Message-ID: <9104111927.AA19652@prodigal.psych.rochester.edu> ANNOUNCEMENT Language Sciences Conference Weekend at the University of Rochester May 9 - 12, 1991 There will be three overlapping conferences: The 1991 CUNY Sentence Processing Conference May 9-11 Workshop on Japanese Linguistics May 10-12 "Belief and Belief Attribution" Philosophy Conference May 11-12 ==> Conference fares are available from USAIR. <== ==> Travel scholarships are available for students. <== We will arrange group vans from nearby cities. **** If you are interested in more information about schedule **** or registration details, please forward a simple YES to: cuny91 at prodigal.psych.rochester.edu **** If you plan to attend ANY events please let us know. **** This is the second of three messages sent to newsgroups; a final conference schedule will be posted next week. From noordewi at cs.rutgers.edu Fri Apr 12 13:00:24 1991 From: noordewi at cs.rutgers.edu (noordewi@cs.rutgers.edu) Date: Fri, 12 Apr 91 13:00:24 EDT Subject: Neural Network Seminar Message-ID: <9104121700.AA08531@porthos.rutgers.edu> RUTGERS UNIVERSITY Dept. of Computer Science/Dept. of Mathematics Neural Networks Colloquium Series --- Spring 1991 C. L. Giles NEC Research Institute Teaching Recurrent Neural Networks to be Finite State Machines (Digraphs) Abstract Recurrent neural networks are natural models for encoding and learning temporal sequences. If these temporal sequences are strings from the languages of formal grammars, then teaching a neural network to learn these sequences is a form of grammatical inference. We demonstrate how to train second-order recurrent networks with real-time learning algorithms to be finite state machines. In particular, we present extensive simulation results which show that simple regular grammars are easy to learn. We devise and use heuristic clustering algorithms which extract finite state machines or digraphs from recurrent neural networks during and after training. The resultant finite state machines usually have large numbers of states and can be reduced in complexity to minimal finite state machines using a standard minimization algorithm. Depending on the training method and type of training set, different minimal finite state machines emerge. If the grammar is well learned, then identical finite state machines are produced in the minimization process. These finite state machines constitute an equivalence class of neural networks which covers different numbers of neurons and different initial conditions. This can be interpreted as a measure of how well a set of strings and its generative grammar are learned. We present a video of the learning process and show the emergent finite state machines during and after training. April 17, 1991 Busch Campus --- 4:30 p.m., room 217 SEC host: Mick Noordewier (201/932-3698) finger noordewi at cs.rutgers.edu for further schedule information From gasser at bend.UCSD.EDU Sat Apr 13 02:13:10 1991 From: gasser at bend.UCSD.EDU (Michael Gasser) Date: Fri, 12 Apr 91 23:13:10 PDT Subject: TR on development of perceptual comparison Message-ID: <9104130613.AA02880@bend.UCSD.EDU> ********************************************************************* DO NOT FORWARD TO OTHER LISTS ********************************************************************* The following paper has been placed in the neuroprose archives at Ohio State University: Comparison, Categorization, and Perceptual Dimensions: A Connectionist Model of the Development of the Notion of Sameness Michael Gasser Linda B. Smith Computer Science Department Psychology Department Indiana University Indiana University Cognitive Science Research Report 41 Abstract The ability to compare two objects is fundamental to intelligent behavior. However, it is important to distinguish between the implicit comparison that plays a role in all categorization and explicit comparison, by which two object representations are compared in short-term memory. Children learn early on both to categorize and to perform explicit comparison, but it requires a long time for them to learn to categorize and compare along particular dimensions. In this paper we present a connectionist model which brings together categorization and comparison, focusing on the development of the use of dimensions in both processes. The model posits a very general comparison mechanism which is blind to the nature of its inputs and sharing of internal object and dimension representations by categorization and comparison processes. As the system is trained on the two processes, it learns to use dimension inputs as filters on the internal representations for input objects; it is these filtered representations which are matched in comparison. The model provides a natural account of the tendency for early comparison along a single dimension to be disrupted by similarities along other, irrelevant dimensions and of the process through which the child might overcome this deficiency. ************************ How to obtain a copy ************************ a) via FTP: % ftp cheops.cis.ohio-state.edu ... Name (cheops.cis.ohio-state.edu:your-id): anonymous ... Password: neuron ... ftp> cd pub/neuroprose ... ftp> binary ... ftp> get gasser.same.ps.Z ... ftp> quit % uncompress gasser.same.ps.Z % lpr -P gasser.same.ps b) via postal mail (only if the above is impossible, please) Request a hardcopy from Beverly Brown, bebbrown at ucs.indiana.edu Cognitive Science Program, Psychology 376 Indiana University Bloomington, IN 47405 USA From tsejnowski at UCSD.EDU Sun Apr 14 01:57:22 1991 From: tsejnowski at UCSD.EDU (Terry Sejnowski) Date: Sat, 13 Apr 91 22:57:22 PDT Subject: Neural Compuation Vol 3 Issue 1 Message-ID: <9104140557.AA23939@sdbio2.UCSD.EDU> NEURAL COMPUTATION - Volume 3 Issue 1 - Spring 1991 Review: Deciphering the Brain's Codes Masakazu Konishi Letters: Synchronization of Bursting Action Potential Discharge in a Model Network of Neocortical Neurons Paul Bush and Rodney Douglas Parallel Activation of Memories in an Oscillatory Neural Network D. Horn and M. Usher Organization of Binocular Pathways: Modeling and Data Related to Rivalry Sidney R. Lehky Dynamics and Formation of Self-Organizing Maps Jun Zhang A Method for Reducing Computation in Networks with Separable Radial Basis Functions Terrence D. Sanger Adaptive Mixtures of Local Experts Robert A. Jacobs, Michael I. Jordan, Steven J. Nowlan, and Geoffrey E. Hinton Efficient Training of Artificial Neural Networks for Autonomous Navigation Dean A. Pomerleau Sequence Manipulation Using Parallel Mapping David S. Touretzky and Deirdre W. Wheeler Parsing Complex Sentences with Structured Connectionist Networks Ajay N. Jain Rules and Variables in Neural Nets Venkat Ajjanagadde and Lokendra Shastri TAG: A Neural Network Model for Large-Scale Optical Implementation Hyuek-Jae Lee, Soo-Young Lee, and Sang-Yung Shin SUBSCRIPTIONS - VOLUME 3 ______ $35 Student ______ $55 Individual ______ $110 Institution Add $18. for postage and handling outside USA (Back issues are available for $28 each.) MIT Press Journals, 55 Hayward Street, Cambridge, MA 02142. (617) 253-2889. MasterCard and VISA accepted ----- From jm2z+ at andrew.cmu.edu Tue Apr 16 09:34:59 1991 From: jm2z+ at andrew.cmu.edu (Javier Movellan) Date: Tue, 16 Apr 91 09:34:59 -0400 (EDT) Subject: Chernoff faces Message-ID: <4c2jq3C00UzxQ1J0VJ@andrew.cmu.edu> Does anybody know of stats packages with a Chernoff faces facility ? This is a method to represent multivariate observations as human faces where each variable controls a particular feature of the face. Thanks Javier Movellan From brp at bandit.berkeley.edu Tue Apr 16 13:52:03 1991 From: brp at bandit.berkeley.edu (Bruce Raoul Parnas) Date: Tue, 16 Apr 91 10:52:03 PDT Subject: models for real neural spike initiators Message-ID: <9104161752.AA00282@bandit.berkeley.edu> I am currently exploring various models used in computer simulations of real neurons, particularly implementations of spike initiators. i have been looking at the two time- constant model proposed by Hill and others many years ago, a temperature-scaled version of the Hodgkin-Huxley model, and the model proposed by Frankenhauser and Dodge. Are people aware of other spike initiator models being used in computer models of biological neurons? thanx, bruce (brp at bandit.berkeley.edu) From wray at ptolemy.arc.nasa.gov Tue Apr 16 21:06:06 1991 From: wray at ptolemy.arc.nasa.gov (Wray Buntine) Date: Tue, 16 Apr 91 18:06:06 PDT Subject: Going to IJCAI-91? Here's a relevant workshop. Message-ID: <9104170106.AA23498@ptolemy.arc.nasa.gov> Some of the nice work on engineering network structures for a given problem, multiple networks, etc., etc., would seem suitable for the following workshop. ----------------------- Call for Papers: IJCAI-91 Workshop Evaluating and Changing Representation in Machine Learning The influence of knowledge representation for the success of machine learning systems is well known. The representation language for a learning application is formed by the vocabulary (signature) as well as by the restrictions on the chosen formalism (e.g.\ number of literals). In practical applications the design of the representation for input, available domain knowledge, expert interaction and output can be time-consuming and critical for success. This workshop will focus on studies that empirically or theoretically evaluate representations, their change, and the construction of new terms. Areas of interest include: logical, relational, probabilistic and neural network representations; concept or classifier learning, model-building, knowledge-intensive learning, theory refinement, and other learning tasks; Bayesian, MDL, PAC or complexity theory; and methodologies for acquisition, evaluation and presentation of different representations and their individual components. The workshop will be held on Saturday, 24 August. Submissions (extended abstracts or short papers, 8-15 pages, 1 copy) should be sent to either of the Program Chairpersons. The collection of accepted papers will be provided to participants, and later collective publication of selected papers will be organised. CHAIRPERSONS Katharina Morik Francesco Bergadano GMD University of Torino GMD - PO Box 1240 Dipartimento di Informatica Schloss Birlinghoven corso Svizzera 185 D-5205 Sankt Augustin 1, Germany Torino, 10149, Italy (+49) 2241 14 2670, fax. 14 2889 (+39) 11 7712002, fax. 11 751603 morik at gmdzi.gmd.de bergadan at di.unito.it Wray Buntine RIACS and AI Research Branch NASA Ames Research Center Mail Stop 244-17 Moffett Field, CA, 94035, USA (+1) 415 6043389, fax. 6046997 wray at ptolemy.arc.nasa.gov IMPORTANT DATES Submission: May 15th Acceptance notified: June 15th Final camera ready copy of full paper: July 15th Workshop: August 24th (Sat.) From gaudiano at park.bu.edu Wed Apr 17 09:52:18 1991 From: gaudiano at park.bu.edu (gaudiano@park.bu.edu) Date: Wed, 17 Apr 91 09:52:18 -0400 Subject: models for real neural spike initiators In-Reply-To: connectionists@c.cs.cmu.edu's message of 16 Apr 91 20:25:12 GM Message-ID: <9104171352.AA06552@copley.bu.edu> >>>>> On 16 Apr 91 20:25:12 GMT, connectionists at c.cs.cmu.edu said: brp> From: Bruce Raoul Parnas brp> I am currently exploring various models used in computer simulations of real neurons, brp> particularly implementations of spike initiators. i have been looking at the two time- brp> constant model proposed by Hill and others many years ago, a temperature-scaled version brp> of the Hodgkin-Huxley model, and the model proposed by Frankenhauser and Dodge. brp> Are people aware of other spike initiator models being used in computer models of brp> biological neurons? A good reference book that covers this topic is: MacGregor, R.J. (1987) "Neural and Brain Modeling". Academic Press. You will find that Chapter 2 is dedicated mostly to the kind of modeling you are interested in, and includes an extensive bibliography. Paolo Gaudiano Department of Cognitive and Neural Systems Boston University 111 Cummington Street Boston, MA 02215 Phone: (617) 353-6181 e-mail: gaudiano at park.bu.edu (ARPANET) From peterc at chaos.cs.brandeis.edu Wed Apr 17 20:23:38 1991 From: peterc at chaos.cs.brandeis.edu (Peter Cariani) Date: Wed, 17 Apr 91 20:23:38 edt Subject: Spike initiation & time-dependent threshold changes In-Reply-To: gaudiano@park.bu.edu's message of Wed, 17 Apr 91 09:52:18 -0400 <9104171352.AA06552@copley.bu.edu> Message-ID: <9104180023.AA28043@chaos.cs.brandeis.edu> Regarding models for spike initiation, as far as I can see there are few (if any) models which take into account the intrinsic time course of threshold following an action potential. Thresholds for many different kinds of axons have a superexcitability period in which the threshold (that amount of injected charge needed to re-excite the membrane) descends to a half or third or less of its resting value. (See Chung, Raymond, & Lettvin, 1970; Raymond & Lettvin, 1978; Raymond, 1979). Furthermore, these thresholds are altered by activity, so the phenomenon of conduction block would serve to dynamically modify the effective connectivity between neurons (in the frequency/interspike interval domain). I don't understand why this phenomenon has been so overlooked by the neuroscience community at large, but I would think it would be extremely important for any models where "intrinsic oscillations" play an important role. I am currently working on an extremely rudimentary model of this sort, trying to understand its functional implications. It's unclear what causes this superexcitable phase--one set of ideas involves electrostatic interaction between calcium and ion channels on the outside of membranes (and is therefore related to Hill's model for accomodation)--but I think it means that neurons have more complex temporal characteristics than has been generally assumed and that oscillatory behavior may not necessarily require recurrent loops with time delays. Obviously this affects the plausibility of temporally-based pulse codes, since what required a reverberatory loop may now only need a single neuron. This all remains to be seen, but it's worth keeping in mind. SH Chung, SA Raymond, & JY Lettvin (1970) Multiple meaning in single visual neurons. Brain Behav Evol 3: 72-101 SA Raymond & JY Lettvin (1978) Aftereffects of activity in peripheral axons as a clue to nervous coding. Physiology and Pathophysiology of Axons S Waxman, ed. Raven Press, NY SA Raymond (1979) Effects of nerve impulses on threshold of frog sciatic nerve fibres. J Physiol 290:273-303. --Peter Cariani From Alexis_Manaster_Ramer at MTS.cc.Wayne.edu Wed Apr 17 22:47:46 1991 From: Alexis_Manaster_Ramer at MTS.cc.Wayne.edu (Alexis_Manaster_Ramer@MTS.cc.Wayne.edu) Date: Wed, 17 Apr 91 22:47:46 EDT Subject: MOL2 notice Message-ID: <321085@MTS.cc.Wayne.edu> ---(Forwarded from: wsavitch at UCSD.EDU, Dated: Wed, 17 Apr 91 20:16:24 EDT)--- From wsavitch at UCSD.EDU Wed Apr 17 20:16:24 1991 From: wsavitch at UCSD.EDU (wsavitch@UCSD.EDU) Date: Wed, 17 Apr 91 17:16:24 PDT Subject: No subject Message-ID: <9104180016.AA11414@walts.ucsd.edu> The Workshop on Mathematics of Language will take place on May 17 and 18 in The Tarrytown Hilton Inn in Tarrytown, New York (just outside of Yorktown Heights, N.Y). The workshop will begin at 9:30 am on May 17 and end at approximately 6:30 pm on May 18. Please note that attendance will be restricted to 80 parti- cipants. Registration will be on a first come, first served basis. To preregister please send a note to MOL2 at WATSON.IBM.COM with the following information: :name. :institution. :e-mail. :phone. If we have not run out of space by then, you will be asked to send in a check for $40 (registration at the door will be $50). All payments will have to be by check. The address of the hotel is: The Tarrytown Hilton Inn 455 South Broadway Tarrytown, New York 10591 The price of a room is $99/night. To reserve a room please call: 914-631-5700 or 1-800-HILTONS and refer to: Association for Computational Linguistics Special Interest Group on the Mathematics of Language The hotel will provide instructions on how to get there. If you have any questions, please contact: Lisa Braden-Harder (MOL 2 -- Local Arrangements) e-mail: MOL2 at WATSON.IBM.COM tel. (914) 784-7849 PROGRAM SCHEDULE FRIDAY, MAY 17, 1991 9:30 A. Manaster-Ramer Opening Address (Capacity, Complexity, and Beyond) 10:00 A. Kornai The Generative Power of Feature Geometry 10:30 BREAK 10:50 J. O'Neil The Generation of the Elementary Tree Set for a Tree-Adjoining Grammar 11:20 Mona Singh The Perfective Aspect: An Algebraic Analysis 11:50 N. Correa Attribute and Unification Grammar: A Review and Analysis of Formalisms 12:20 LUNCH 1:45 D. Radzinski Chinese Distributive Numerals and Indexed Grammars 2:15 Munidar Singh Towards a Formal Semantics for Speech Acts 2:45 BREAK 3:15 M. A. Moshier On Completeness Theorems for Feature Logics 3:45 E. Ristad On the Complexity of Two Problems in Human Language 4:15 BREAK 4:35 A. Joshi and Y. Schabes Fixed and Flexible Phrase Structure: Coordination in Tree Adjoining Grammars 5:05 DINNER 7:30 M. Dymetman A Generalized Greibach Normal Form for Definite Clause Grammars and the Decidability of the Offline-Parsability Problem 8:00 J. Nerbonne A Feature-Based Syntax/Semantics Interface 8:30 D. Johnson & L. Moss Languages Generated by Multistratal Axiomatic Grammars SATURDAY, MAY 18, 1991 9:30 R. Berwick Asymptotic Computational Complexity and Average Case Computational Complexity for Natural Language Analysis 10:00 R. Zuber An Algebraic Approach to Negation 10:30 BREAK 10.50 S. Zeitman Somewhat Finite Approaches to Infinite Sentences From tgd at turing.CS.ORST.EDU Thu Apr 18 00:14:11 1991 From: tgd at turing.CS.ORST.EDU (Tom Dietterich) Date: Wed, 17 Apr 91 21:14:11 PDT Subject: Preprint: Error-correcting output codes Message-ID: <9104180414.AA12756@turing.CS.ORST.EDU> ** Please do not forward to other newsgroups ** The following preprint is available for FTP access from the Ohio State Neuroprose archive: Error-Correcting Output Codes: A General Method for Improving Multiclass Inductive Learning Programs Thomas G. Dietterich and Ghulum Bakiri Department of Computer Science Oregon State University Corvallis, OR 97331-3202 Multiclass learning problems involve finding a definition for an unknown function f(x) whose range is a discrete set containing k>2 values (i.e., k ``classes''). The definition is acquired by studying large collections of training examples of the form . Existing approaches to this problem include (a) direct application of multiclass algorithms such as the decision-tree algorithms ID3 and CART, (b) application of binary concept learning algorithms to learn individual binary functions for each of the k classes, and (c) application of binary concept learning algorithms with distributed output codes such as those employed by Sejnowski and Rosenberg in the NETtalk system. This paper compares these three approaches to a new technique in which BCH error-correcting codes are employed as a distributed output representation. We show that these error-correcting output representations improve the performance of ID3 on the NETtalk task and of backpropagation on an isolated-letter speech-recognition task. These results demonstrate that error-correcting output codes provide a general-purpose method for improving the performance of inductive learning programs on multiclass problems. This paper will appear in AAAI-91. To retrieve: unix> ftp cheops.cis.ohio-state.edu (or 128.146.8.62) Name: anonymous Password: your-name at your-address ftp> cd pub/neuroprose ftp> binary ftp> get dietterich.error-correcting.ps.Z ftp> bye unix> uncompress dietterich.error-correcting.ps.Z unix> lpr -P{postscript printer} dietterich.error-correcting.ps Thank-you to Jordan Pollack for providing the ftp services. --Tom From brp at bandit.berkeley.edu Thu Apr 18 12:10:55 1991 From: brp at bandit.berkeley.edu (Bruce Raoul Parnas) Date: Thu, 18 Apr 91 09:10:55 PDT Subject: references to spike initiators Message-ID: <9104181610.AA02371@bandit.berkeley.edu> there have been several requests for references to the spike initiator models i mentioned in a previous message, so, here they are: One of the oldest models, and still used by many (including Macgregor) as the basis for a great many models, is the system derived by A.V. Hill, the two-time-constant model: A.V. Hill (1936), "Excitation and accomodation in nerve", Proc. Royal Soc., Series B, 119: 305-353. The model of Frankenhaeuser and Dodge (actually it'd Dodge's model, modified by Frankenhaeuser and Huxley) was developed for the frog node of Ranvier: Frankenhaeuser, B. and Huxley, A.F. (1964) "The action potential in the myelinated nerve fiber of Xenopus Laevis as computed on the basis of voltage clamp data", J. Physiol. 171: 302-315. (note: the spelling may be either Frankenhaeuser or frankenhauser, depending on where you look, the ae is used to indicate an a-umlaut in the German spelling). And, of course, there is the HH model, presented in a series of papers: Hodgkin, A.E. and Huxley, A.F. in J. Physiol., 1951 and 1952. bruce (brp at bandit.berkeley.edu) From chan%unb.ca at UNBMVS1.csd.unb.ca Thu Apr 18 16:08:27 1991 From: chan%unb.ca at UNBMVS1.csd.unb.ca (Tony Chan) Date: Thu, 18 Apr 91 17:08:27 ADT Subject: Call for participation in AI Symposium Message-ID: *************************** * * * CALL for PARTICIPATION * * * *************************** The Fourth UNB Artificial Intelligence Symposium University of New Brunswick Fredericton, New Brunswick, Canada Friday & Saturday, Sept. 20 & 21, 1991 Program Committee: ****************** Ranan Banerji, St. Joseph's U., Philadelphia, PA Wolfgang Bibel, I.T.H., Darmstadt, West Germany David Bonham, Mechanical Eng., U.N.B. Z. Chen, U. of Nebraska, Omaha, NE Chang Choo, W.P.I., Worcester, MA Rajamani Doraiswami, Electrical Eng., U.N.B. Martin A. Fischler, SRI Int., Menlo Park, CA Herbert Freeman, Rutgers U., New Brunswick , NJ C. Lee Giles, NEC Research Ins., Princeton, NJ Paul Gillard, Memorial U., St. John's, Nfld. Vasant Honavar, Iowa State U., Ames, IA Laveen Kanal, U. of Maryland, Coll. Park, MD Vladik Kreinovich, U. of Texas, El Paso, TX Werner Kuhn, U. of Maine, Orono, ME Bernd Kurz, Computer Science, U.N.B. Patrice Lapointe, AECL Res., Chalk River, Ont. Mark Lidd, Mitre Corp., Fairfax, VA B.I.B. Madhav, I.I.T., Madras, India Ettore Merlo, C.R.I.M., Montreal, Que. Tom Mitchell, Carnegie Mellon, Pittsburgh, PA Jan Mulder, Dalhousie U., Halifax, N.S. Eric Neufeld, U. of Saskatchewan, Saskatoon, Sask. N. Parameswaran, I.I.T., Madras, India Richard Peacocke, Bell-Northern Res., Ottawa, Ont. Denis Riordan, T.U.N.S., Halifax, N.S. John Robinson, Civil Eng., U.N.B. Azriel Rosenfeld, U. of Maryland, Coll. Park, MD Leemseop Shim, Governors State U., Univ. Park, IL Eduardo Sontag, Rutgers U., New Brunswick, NJ Paul Tarau, U. de Moncton, Moncton, N.B. Manoel Tenorio, Purdue U., West Lafayette, IN Richard Tervo, Electrical Eng., U.N.B. Tetsuyuki Toyofuku, National-Panasonic, Osaka, Japan Andri Trudel, Acadia U., Wolfville, N.S. Lloyd Waugh, Civil Eng., U.N.B. Ian Witten , U. of Calgary, Calgary, Alta. Andrew K.C. Wong, U. of Waterloo, Waterloo, Ont. Jean Zanazaka, U. de Sherbrooke, Que. Sponsored by: ************* Faculty of Computer Science, University of New Brunswick Bell-Northern Research Canadian Society for Computational Studies of Intelligence In Cooperation with: ******************** The American Association for Artificial Intelligence (AAAI) IEEE Computer Society International Association of Knowledge Engineers (IAKE) Japanese Society of Artificial Intelligence Keynote Address: **************** "Self-Reliant Robots: The Ambler Rover and Beyond" by Dr. Reid Simmons, Research Computer Scientist, Robotics Institute, Carnegie Mellon University, Pittsburgh, PA Panel Discussion (1.5 hours): ***************************** "On the Role of Machine Learning in Artificial Intelligence" ---------------------------------------------------------------------- The UNB Artificial Intelligence Symposium provides a forum for an exchange of ideas, experiences and information about activities in artificial intelligence. Your participation is invited in the form of (1) a paper to be reviewed (nominally by three reviewers) and, if accepted, published and presented at the Symposium, or (2) a half-day tutorial on a suitable AI topic, to be presented at the Symposium. All tutorial proposals must have a stated goal, a detailed outline (2 - 5 pages), target audience defined, and a brief resume of the tutorial presenter(s), or (3) an exhibit or display of artificial intelligence technology. Important Dates: **************** May 15, 1991 -- Four copies of an extended abstract (2 to 4 double- spaced pages) or a full paper (max. 11 pages, including figures) to one of the Program Co-Chairs (see address below). July 1, 1991 -- Notification of acceptance will be mailed. August 15, 1991 -- Final camera-ready copies of papers are due. Topics of interest include, but are not limited to *************************************************** Foundations of AI, Connectionist Theory & Applications Pattern Recognition, Machine Learning Knowledge Acquisition, Automated Theorem Proving Computer Vision, Knowledge Representation Natural Language Processing , Temporal Reasoning Knowledge-Based Systems, Database/Knowledge Base Integration Robotics, Real-Time Expert Systems, Automated Planning Expert System Verification, Automated Scheduling Novel Architectures for AI, Case-Based Reasoning Truth Maintenance Systems, Reasoning Under Uncertainty A pre-published proceedings will be available at the Symposium. Submit papers or extended abstracts to: Program Co-Chairs, 4th UNB AI Symposium Brad Nickerson or Lev Goldfarb E-mail: bgn at unb.ca or goldfarb at unb.ca Submit tutorial proposals to: Bruce Spencer or Przemyslaw Pochec Tutorial Co-Chairs, 4th UNB AI Symposium E-mail: bspencer at unb.ca or pochec at unb.ca Submit exhibit requests to: Kirby Ward Exhibits Chair, 4th UNB AI Symposium E-mail: wardk at unb.ca The mailing address, phone and fax numbers for the above are as follows: Faculty of Computer Science, University of New Brunswick P.O. Box 4400, Fredericton, N.B., Canada E3B 5A3 Phone: (506) 453-4566 Fax: (506) 453-3566 It is planned to hold the CKEd (Certified Knowledge Engineer) examination on Sept. 19, 1991 at the Symposium location. Symposium participants wishing to take the examination should contact the IAKE at International Associaton of Knowledge Engineers Georgetown P.O. Box 25461, Washington, D.C. 20007, U.S.A. ph: 301 231 7826 fax: 301 770 4621 E-mail: IAKE at UC780.bitnet From koch at CitIago.Bitnet Thu Apr 18 16:59:13 1991 From: koch at CitIago.Bitnet (Christof Koch) Date: Thu, 18 Apr 91 13:59:13 PDT Subject: Spike initiation & time-dependent threshold changes In-Reply-To: Your message <9104180023.AA28043@chaos.cs.brandeis.edu> dated 17-Apr-1991 Message-ID: <910418135913.2040bf1d@Iago.Caltech.Edu> The fact that the threshold of firing goes up following action potential dischar ge is a well know and well studied model. It's paartially caused by the influx of calcium, which activates one or two calcium-dependent potassium currents. See the model in Chapter 4 ("Multiple channels and calcium dynamics" in the book "Methods in Neuronal Modeling", (C. Koch and I. Segev, eds.), MIT Press, 1989. This book contains a number of quite detailed, state-of-the-art models. One of the best studied models of spike-initiation is still the integrate-and-fire or leaky-integrate-and-fire models, develoepd by Brice Knight. It's quite well described in Tuckwell's two volume monograph "Introduction to theoretical neurobiology" (Cambridge Univ. Press, 1988). Some analytical solutions are known, for instance in the case that the input is a Poisson input. Christof From brp at bandit.berkeley.edu Fri Apr 19 00:44:19 1991 From: brp at bandit.berkeley.edu (Bruce Raoul Parnas) Date: Thu, 18 Apr 91 21:44:19 PDT Subject: Spike initiation & time-dependent threshold changes Message-ID: <9104190444.AA03832@bandit.berkeley.edu> One of the best studied models of spike-initiation is still the integrate-and-fire or leaky-integrate-and-fire models, develoepd by Brice ^^^^^ Knight. It's quite well described in Tuckwell's two volume monograph ^^^^^^ "Introduction to theoretical neurobiology" (Cambridge Univ. Press, 1988). Christof While i agree that simple integrate and fire models (one time constant or two) are definitely among the best around, I find it odd thatt so many people attribute these models to people other that those who developed them. The earliest work I know of on two time constant models was done in 1936 by A.V. Hill, when much less was known about neurophysiology. Rashevsky and Monier worked on similar models at about the same time, but the concept is 55 years old. bruce From cuny91 at prodigal.psych.rochester.edu Fri Apr 19 08:36:05 1991 From: cuny91 at prodigal.psych.rochester.edu (A. Zukowski (CUNY Coordinator)) Date: Fri, 19 Apr 91 08:36:05 EDT Subject: No subject Message-ID: <9104191236.AA11419@prodigal.psych.rochester.edu> PLEASE POST HARD COPY WHERE PERTINENT LANGUAGE SCIENCES WEEKEND AT THE UNIVERSITY OF ROCHESTER THERE WILL BE THREE CONFERENCES ON LANGUAGE MAY 9-12 AT THE UNIVERSITY OF ROCHESTER. ANNUAL CUNY SENTENCE PROCESSING CONFERENCE ANNUAL JAPANESE GRAMMAR CONFERENCE ROCHESTER PHILOSOPHY CONFERENCE ON BELIEF AND BELIEF ATTRIBUTION WE HAVE CRASH SITES, TRAVEL FELLOWSHIPS AND OTHER LOCAL ARRANGEMENTS AVAILABLE. IF YOU ARE INTERESTED IN ATTENDING, PLEASE CONTACT US: NAME: EMAIL: PHONE: SURFACE MAIL: SCHEDULES FOLLOW 1991 CUNY Sentence Processing Conference Schedule ------------------------------------------------- THURSDAY, MAY 9 12:55 -1:00 Opening Remarks 1:00 - 1:35 Ivan Sag "Taking Performance Seriously: A Plea for Constraint" 1:35 - 2:00 Weijia Ni & Stephen Crain "Avoiding Garden Paths" 2:00 - 2:25 Gerry Altmann, Alan Garnham, Yvette Dennis, & A. Henstra "Eye Movements Controlled: Evidence For and Against Context Effects" 2:25 - 2:50 Don Mitchell, Martin Corley, & Alan Garnham "Biases and Influences in Human Sentence Parsing: Evidence for the Precedence of "Pure" Syntactic Strategies" 2:50 - 3:15 Maryellen C. MacDonald "Pre- and Post-ambiguity Cues for Syntactic Ambiguity Resolution" 3:15 - 4:00 Break 4:00 - 5:30 Japanese Psycholinguistics Symposium (Reiko Mazuka, Mineharu Nakayama, Yoshi Otsu, Peter Culicover) 5:50 - 7:30 Supper 7:30 - 9:30 Connectionism and Models of Language (Paul Smolensky, Gary Dell, Jerry Feldman) ------------------------------------------------------------------------------- FRIDAY, MAY 10 8:30 - 9:00 Coffee 9:00 - 9:25 Janet Nicol, Cecile McKee, & Dana McDaniel "The Development of the Coreference Processor" 9:25 - 9:50 Howard Kurtzman, Raul Elias-Cintron, & Rocio Aramburu "Probe Recognition in Spanish: Effects of NP-trace and Agreement Morphology" 9:50 - 10:15 Karen Emmorey & Diane Lillo-Martin "Processing Spatial Anaphora: Referent Activation with Overt and Null Pronouns in American Sign Language" 10:15 - 10:40 Kay Bock & Kathleen M. Eberhard "The Meaning, Sound, and Syntax of English Number Agreement" 10:40 - 11:00 Break 11:00 - 12:30 Second Language Acquisition Symposium (Peter Jusczyk, Jacques Mehler, Elissa Newport) 12:30 - 2:00 Lunch 2:00 - 2:25 Di Bradley "Lexical and Post-Lexical Contributions to Priming" 2:25 - 2:50 (To Be Announced) 2:50 - 3:15 Michael Brent "Automatic Acquisition of Subcategorization Frames from Untagged, Free-Text Corpora" 3:15 - 3:40 Mary Potter "Word Perception in Sentence Context: Back to the Future" 3:40 - 4:00 Break 4:00 - 5:30 Conditionals Symposium (William Lycan, Phillip Johnson-Laird) 5:30 - 7:30 POSTERS Tod A. Bassham & Jill N. Beckman "The Parsing of Tough- and Too-Constructions: Implications for Models of Gap-Location" Tom Bever, Cornell Juliano, & Andrea Zukowski "Improvement of Text Readability Through Applied Psycholinguistics" Elizabeth R. Blackmer "Implications of the Timing of Rapid Multiple Repairs in Spontaneous Speech" Giuseppe Cappelli, Anna Maccari & Lucia Pfanner "A System for Semi-Automatical Treatment of Child Morphology" Soon Ae Chun "Verbal Suffixes and Control Interpretation in Korean" Mark Coulson "Processing Ambiguous Anaphora" Aaron Halpern "Priming in VP Anaphora" Gregory Hickok, Enriqueta Canseco, Edgar Zurif, & Jane Grimshaw "Modularity in Locating Wh-gaps" Dan Jurafsky "An On-Line Model of Human Sentence Interpretation" Sheila Meltzer "(How) Does Discourse Pragmatics Influence Processing of Phonologically Null Pronouns?" Gail Mauner "Syntactic Context and the Interpretation of Verb Phrase Anaphors" Neal Perlmutter "Effects of Pragmatic Cues on Syntactic Ambiguity Processing" Justin Peterson & Dorrit Billman "I'm with Her and the Butler Did It with the Knife" Susan M. Powers "'I Give Up': Children's Processing of Sentences Containing Verb Particles" Leslie Re, Harriet Taber, & Janet D. Fodor "Reconciling Methodological Issues in Empty Category Processing" Vincent J. Samar "Be is a Raising Verb: Psycholinguistic Evidence" Yoshinori Sasaki "Development of Sentence Processing Strategies in English and Japanese as Foreign Languages: An Analysis Based on the Competition Model" Jan C. Scholtes "Kohonen's Self Organizing Map Applied Towards Natural Language Processing" Julie Sedivy "The Use of Thematic Relations in Gap-Filling" L. P. Shapiro, H. N. Nagel, & B. A. Levine "Verb Frame Preferences and Sentence Processing: Implications for a Model of Parsing" Ron Smyth & Kumiko Murasugi "Children's Use of Agreement Clues to Relative Clause Attachment" Laurie A. Stowe, Peter W. Culicover, Michael Torello, Robert Angel, & Tod Bassham 7:30 - 9:30 Banquet 9:30 - Festivities ------------------------------------------------------------------------------- SATURDAY, MAY 11 9:00 - 9:30 Coffee 9:30 - 10:15 Susan Garnsey - ERP Tutorial 10:15 - 10:40 Rob Kluender & Marta Kutas "Evidence for Lexical and Syntactic Processing Effects in Wh-Islands" 10:40 - 11:00 Break 11:00 - 1:00 Reading Symposium (Marcel Just, Mark Seidenberg, Gail McKoon) 1:00 - 2:00 Lunch 2:00 - 2:25 Janet Dean Fodor "Empty Categories at S-Structure and PF" 2:25 - 2:50 Andrew Barss "The Grammar of Antecedence and Antecedent Reactivation" 2:50 - 3:15 Martin Pickering & Guy Barry "Gap-Free Processing and the Resolution of Ambiguities" 3:15 - 4:00 Edward Gibson "A Computational Treatment of Processing Overload" 4:00 - 4:05 Break 4:05 - 5:35 Lexical Concepts Symposium (Frank Keil, James Hampton) ********************** The Rochester Workshop on Japanese Linguistics, Universal Grammar, and Their Implications to Language Pedagogy and Human Cognition (tentative schedule) --------------------------------------------------------------- Friday, May 10 -------------- 9.00-9.15 Opening Remarks 9.15-10.15 On Scrambling Naoko Nemoto, University of Connecticut 10.15-10.30 Break 10.30-11.30 Scrambling and A-chains Shigeru Miyagawa, Ohio State University 11.30-12.30 Scrambling and the FOCUS Interpretation Ayumi Ueyama, Kyoto University of Foreign Studies 12.30-2.15 Lunch 2.15-3.15 Focused Pronouns Richard Larson, SUNY Stony Brook 3.15-4.15 Subject-to-Object Raising, ECM and the Major Object in Japanese Hojime Hoji, USC 4.15-4.30 Break 4.30-5.30 The Syntax of Answers: A Comparative Approach Itziar Laka, University of Rochester 7.30-9.30 Banquet (with CUNY Conference participants) Saturday, May 11 ---------------- 9.45-10.45 Reanalyzing Reanalysis Natsuko Tsujimura, University of Illinois 10.45-11.45 Title TBA Shige-Yuki Kuroda, UC San Diego 11.45-2.00 Lunch 2.00-3.00 On Passive Hiroto Hoshi, University of Connecticut 3.00-4.00 De-Passivization Yoshihisa Kitagawa, University of Rochester 4.00-4.15 Break 4.15-5.15 Two Types of Inflection David Pesetsky, MIT Sunday, May 12 -------------- 10.00-11.00 Negative Polarity Items and Floating Quantifiers Naoya Fujita, University of Rochester 11.00-12.00 Dissussion Session 1) Comments on Nemoto, Miyagawa, and Ueyama Hiroaki Tada, MIT 2) Comments on Larson Shin Watanabe,USC ***************************** The Univerisity of Rochester Department of Philosphy's Conference on Belief and Belief Attribution May 11-12, 1991 --------------------------------------------------------------------- Saturday, May 11 ---------------- 9:30 Robert Stalnaker "On Twin Earth" 11:00 Ernest LePore "Holism & Belief" --lunch-- 2:00 William Lycan "A New Argument for the Representational Theory of Thinking" 3:00 Mark Crimmins "Tacitness Doesn't Make a Difference Sunday, May 12 -------------- 9:30 James Higginbotham "Belief & Logical Form" 11:00 Scott Soames "Attitudes & Anaphora" Information about the conferences may be requested from the following sources: CUNY: Andrea Zukowski (cuny91 at prodigal.psych.rochester.edu) Psychology Department University of Rochester Rochester, NY 14627 (716) 275-0754 (716) 275-8724 Japanese Linguistics, UG Workshop: Yoshi Kitagawa (ykit%uorvm.bitnet at db1.cc.rochester.edu) Foreign Languages, Literature, & Linguistics University of Rochester Rochester, NY 14627 (716) 275-4298 (716) 275-4252 Belief and Belief Attribution Conference: David Braun (dbrn at uhura.cc.rochester.edu) Department of Philosophy Univerisity of Rochester Rochester, NY 14627 (716) 275-4105 From stolcke at ICSI.Berkeley.EDU Sat Apr 20 18:27:39 1991 From: stolcke at ICSI.Berkeley.EDU (stolcke@ICSI.Berkeley.EDU) Date: Sat, 20 Apr 91 16:27:39 MDT Subject: new neta release of cluster program Message-ID: <9104202327.AA16136@icsib30.Berkeley.EDU> Dear Connectionists: This is the second release of my enhanced version of the `cluster' program (original version by Y. Miyata). It has a bunch of added features, most importantly I merged a Principal Component Analysis program I once wrote into this program, so it now does both HCA and PCA, using compatible input and output formats. I and a couple of people around here have been using this with great success, but there are bound to be some bugs left, especially in the less commonly used features. I would encourage people who have a need for this kind of tool to give it a try and report errors and shortcomings back to me. Once I'm confident that there are no major deficiencies left I'll make an updated version available and announce it on this list. Andreas HOW TO GET CLUSTER cluster is available via anonymous ftp from icsi-ftp.berkeley.edu (128.32.201.55). To get it use FTP as follows: % ftp icsi-ftp.berkeley.edu Connected to icsic.Berkeley.EDU. 220 icsi-ftp (icsic) FTP server (Version 5.60 local) ready. Name (icsic.Berkeley.EDU:stolcke): anonymous Password (icsic.Berkeley.EDU:anonymous): 331 Guest login ok, send ident as password. 230 Guest login Ok, access restrictions apply. ftp> cd pub/ai 250 CWD command successful. ftp> binary 200 Type set to I. ftp> get cluster-2beta.tar.Z 200 PORT command successful. 150 Opening BINARY mode data connection for cluster-2beta.tar.Z (15531 bytes). 226 Transfer complete. 15531 bytes received in 0.08 seconds (1.9e+02 Kbytes/s) ftp> quit 221 Goodbye. HOW TO BUILD CLUSTER Unpack in an empty directory using % zcat cluster-2beta.tar.Z | tar xf - Read the README and especially the man page (cluster.man) for information. Then compile with % make After making the appropriate adjustments in the Makefile you can % make install On most BSD-based UNIX systems cluster should compile without problems. The Makefile gives some hint as to what might have to be changed to get it to run. And please, tell me where I goofed! From tsejnowski at UCSD.EDU Sat Apr 20 22:32:42 1991 From: tsejnowski at UCSD.EDU (Terry Sejnowski) Date: Sat, 20 Apr 91 19:32:42 PDT Subject: Neural Computation Long Papers Message-ID: <9104210232.AA26247@sdbio2.UCSD.EDU> The editorial board of Neural Computation has decided to start accepting a limited number of full length papers. These will be more thoroughly reviewed than letters and there will be only one or two per issue, so it will be more difficult to get one accepted, like the longer articles in Nature and Science. Terry ----- From gelenbe at csd36.NYU.EDU Mon Apr 22 12:26:56 1991 From: gelenbe at csd36.NYU.EDU (Erol Gelenbe) Date: Mon, 22 Apr 91 12:26:56 -0400 Subject: IEEE Transactions on Software Eng. Special Issue Message-ID: <9104221626.AA05267@csd36.NYU.EDU> A special isssue of the IEEE TSE on "Artificial Neural Network Models and Systems" is planned for the month of April 1992, for which I will as Editor. Papers concerning models and their theory, system implementations including one or more applications, and system environments which use artificial neural networks are sollicited. They will be reviewed in accordance with the normal practices of these IEEE Transactions. Paper submissions should be sent to me by August 1, 1991 to the following address : Erol Gelenbe EHEI 45 rue des Saints-Peres 75006 Paris, France erol at ehei.ehei.fr From fay at archsci.arch.su.OZ.AU Tue Apr 23 05:49:34 1991 From: fay at archsci.arch.su.OZ.AU (fay@archsci.arch.su.OZ.AU) Date: Tue, 23 Apr 91 19:49:34 +1000 Subject: Symposium on AI, Reasoning and Creativity Message-ID: <9104230949.AA09174@chomsky.arch.su.OZ.AU> Could you please let me know if you are interested in attending or submitting a paper to this symposium. Thanks Fay * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * PRELIMINARY ANNOUNCEMENT AND CALL FOR PAPERS * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * SYMPOSIUM ON ARTIFICIAL INTELLIGENCE, REASONING AND CREATIVITY 20-23 AUGUST 1991 immediately preceding the International Joint Conference on Artificial Intelligence (IJCAI'91) organised by GRIFFITH UNIVERSITY, BRISBANE, AUSTRALIA to be held at LAMINGTON NATIONAL PARK, QUEENSLAND, AUSTRALIA AIM Creativity is one of the least understood aspects of intelligence. It is commonly regarded as 'intuitive' and not susceptible to rational enquiry. However, there is now considerable work in artificial intelligence and cognitive science which addresses creativity. This symposium will provide a forum for exploring and discussing these ideas, and for suggesting directions for future research. It aims to attract practitioners of both 'cognitive' and 'technological' artificial intelligence. KEYNOTE SPEAKER - PROFESSOR MARGARET BODEN Margaret Boden, in her talk on `Creativity and Computers', will discuss how computational concepts drawn from artificial intelligence can explore creativity. Computers can sometimes do apparently creative things; more to the point, they can suggest how we manage to do so. Computational ideas are therefore helping us to understand how human originality is possible. Margaret Boden is Professor of Philosophy and Psychology, and Founding Dean of the School of Cognitive Sciences, at the University of Sussex, UK. Her recent publications include `Artificial Intelligence and Natural Man' (1987), `Artificial Intelligence in Psychology' (1989) and `The Creative Mind: Myths and Mechanisms (1991). INVITED SPEAKERS include Ernest Edmonds, Loughborough University, UK John Gero, University of Sydney, Australia Graham Priest, University of Queensland, Australia Roger Wales, University of Melbourne, Australia TOPICS FOR PRESENTATION AND DISCUSSION The symposium calls for extended abstracts of 750-1,000 words. The following are suggested areas only: - Models of creativity - Modelling creative processes - Creative reasoning, e.g theory generation in science - Analogical reasoning - Case-based reasoning - Nonmonotonic reasoning - Thought experiments Both connectionist and traditional symbolic approaches are welcome. All abstracts will be refereed. Accepted papers will be subjected to a further refereeing process for publication by Kluwer Academic Press. The symposium will be structured to provide adequate time for both presentation and discussion. SUBMISSION Three copies of extended abstracts are required by 31 May 1991. Abstracts may be submitted electronically as LaTeX or plain ASCII files via email but hard copies must also be submitted. Two hard copies of final versions of accepted papers, and an electronic version on Macintosh disk or via email are required at the time of registration on 20 August 1991. Authors of accepted papers will be expected to sign a copyright release form to allow publication of the proceedings. At least one author of each paper is expected to present the paper at the symposium. PREPRINTS AND PROCEEDINGS Accepted extended abstracts will be printed in the form of Preprints and be available for distribution at the time of registration. Full papers will be published subsequently by Kluwer Academic Press. LOCATION The symposium is being held at O'Reillys Lodge in Lamington National Park, Queensland - a rainforest 120 km south of Brisbane. Rainforests vary from the temperate beech forest of the higher altitudes to the warm subtropical rainforest of the valleys. There is an abundance of rare and spectacular plants (orchids, ferns, giant epiphytes, mosses, luminous fungi) and a diverse community of birds and animals (crimson rosellas, king parrots, bower birds, miniature kangaroos, brush turkeys) all coexisting beneath the lofty rainforest canopy. On the final day of the symposium there will be ample time for guided or independent bush walks, 4WD bus trips, barbeques and evening entertainment. FACILITIES O'Reillys is a mountain resort of modern units. The seminar facilities are highly professional, including lecture theatre, audiovisual equipment and library. TIMETABLE Extended abstracts (750-1000 words) - 3 hard copies 31 May 1991 Notification of acceptances 17 June 1991 Full formatted papers due 20 August 1991 Symposium 20-23 August 1991 COSTS in Australian dollars (US$1 = ~A$1.28; PStg 1 = ~A$2.33) Registration fee (including one copy of Preprints): Full fee $250 Authors (1 per paper) $150 Accommodation (including all meals): Bethongabel units (private bath, balcony, view) $119 pp/pn Elabana units (private bath, limited availability) $105 pp/pn Bus to Lamington National Park from Griffith University: Round trip $25 INTERNATIONAL COMMITTEE Chair Terry Dartnall, Griffith University, Australia Conference Organiser Fay Sudweeks, University of Sydney, Australia Advisory Board Margaret Boden, Sussex University, UK Andy Clarke, Sussex University, UK Max Coltheart, Macquarie University, Australia Boi Faltings, Ecole Polytechnique Federale de Lausanne, Switzerland Marilyn Ford, Griffith University, Australia John Gero, University of Sydney, Australia Rod Girle, Griffith University; Australian National University Laveen Kanal, University of Maryland, USA Graham Priest, University of Queensland, Australia Simon Ross, University College of London, UK; Kluwer Academic Press Aaron Sloman, Sussex University, UK Roger Wales, University of Melbourne, Australia Janet Wiles, University of Queensland, Australia CONFERENCE CONTACTS Correspondence and queries: Dr Terry Dartnall School of Computing and Information Technology Griffith University Nathan Qld 4111 Australia Tel: +61-7-875 5020 Fax: +61-7-875 5198 Email: terryd at gucis.sct.gu.edu.au Abstracts and papers: Ms Fay Sudweeks Department of Architectural and Design Science University of Sydney NSW 2006 Australia Tel: +61-2-692 2328 Fax: +61-2-692 3031 Email: fay at archsci.arch.su.oz.au Registration: Ms Denise Vercoe School of Computing and Information Technology Griffith University Nathan Qld 4111 Australia Tel: +61-7-875 5002 Fax: +61-7-875 5198 From collins at z.ils.nwu.edu Tue Apr 23 12:58:23 1991 From: collins at z.ils.nwu.edu (Gregg Collins) Date: Tue, 23 Apr 91 11:58:23 CDT Subject: On-campus housing for ML-91 Message-ID: <9104231658.AA00525@z.ils.nwu.edu> We are pleased to announce the availability of on-campus housing for ML91, the Eighth International Workshop on Machine Learning, which will be held at Northwestern University, Evanston, Illinois, USA, June 27-29, 1991. Housing will consist of double-occupancy dormitory rooms at a rate of $69.00 for four nights (June 26-29). *****************ML91 On-Campus Housing Registration Form***************** Please send this form, along with a check for $69.00 made out to Northwestern University, to the following address: Machine Learning 1991 The Institute for the Learning Sciences 1890 Maple Avenue Evanston, Illinois, 60201 USA phone (708) 491-3500 fax (708) 491-5258 email ml91 at ils.nwu.edu Registration information (please type or print): Name: Sex: Address: Phone: Email: Name of person with whom you would like to share your room (if left blank, we will assign you a roommate): From collins%z.ils.nwu.edu at VM.TCS.Tulane.EDU Tue Apr 23 12:58:23 1991 From: collins%z.ils.nwu.edu at VM.TCS.Tulane.EDU (Gregg Collins) Date: Tue, 23 Apr 91 11:58:23 CDT Subject: On-campus housing for ML-91 Message-ID: <9104231658.AA00525@z.ils.nwu.edu> We are pleased to announce the availability of on-campus housing for ML91, the Eighth International Workshop on Machine Learning, which will be held at Northwestern University, Evanston, Illinois, USA, June 27-29, 1991. Housing will consist of double-occupancy dormitory rooms at a rate of $69.00 for four nights (June 26-29). *****************ML91 On-Campus Housing Registration Form***************** Please send this form, along with a check for $69.00 made out to Northwestern University, to the following address: Machine Learning 1991 The Institute for the Learning Sciences 1890 Maple Avenue Evanston, Illinois, 60201 USA phone (708) 491-3500 fax (708) 491-5258 email ml91 at ils.nwu.edu Registration information (please type or print): Name: Sex: Address: Phone: Email: Name of person with whom you would like to share your room (if left blank, we will assign you a roommate): From harnad at Princeton.EDU Tue Apr 23 16:31:25 1991 From: harnad at Princeton.EDU (Stevan Harnad) Date: Tue, 23 Apr 91 16:31:25 EDT Subject: 2 TRs: Categorical Perception and Neural Nets Message-ID: <9104232031.AA00250@reason.Princeton.EDU> The following two tech reports are available by anonymous ftp from directory /pub/harnad on princeton.edu. Full ftp instructions follow the abstracts. ------------------------------------------------- (1) Categorical Perception and the Evolution of Supervised Learning in Neural Nets S Harnad*, SJ Hanson*,** & J Lubin* *Princeton University **Siemens Research Center [Presented at 1991 AAAI Symposium on Symbol Grounding: Problem and Practice] ABSTRACT: Some of the features of animal and human categorical perception (CP) for color, pitch and speech are exhibited by neural net simulations of CP with one-dimensional inputs: When a backprop net is trained to discriminate and then categorize a set of stimuli, the second task is accomplished by "warping" the similarity space (compressing within-category distances and expanding between-category distances). This natural side-effect also occurs in humans and animals. Such CP categories, consisting of named, bounded regions of similarity space, may be the ground level out of which higher-order categories are constructed; nets are one possible candidate for the mechanism that learns the sensorimotor invariants that connect arbitrary names (elementary symbols?) to the nonarbitrary shapes of objects. This paper examines how and why such compression/expansion effects occur in neural nets. [Retrieve by anonymous ftp in binary mode as (compressed) file harnad91.cpnets.Z from directory /pub/harnad on princeton.edu, instructions below] ----------------------------------------------------------------- (2) Connecting Object to Symbol in Modeling Cognition Stevan Harnad Department of Psychology Princeton University Princeton NJ 08544 [To appear in Clark, A. & Lutz, R. (Eds) (1992) "CONNECTIONISM IN CONTEXT," Springer-Verlag] Connectionism and computationalism are currently vying for hegemony in cognitive modeling. At first glance the opposition seems incoherent, because connectionism is itself computational, but the form of computationalism that has been the prime candidate for encoding the "language of thought" has been symbolic computationalism, whereas connectionism is nonsymbolic. This paper examines what is and is not a symbol system. A hybrid nonsymbolic/symbolic system will be sketched in which the meanings of the symbols are grounded bottom-up in the system's capacity to discriminate and identify the objects they refer to. Neural nets are one possible mechanism for learning the invariants in the analog sensory projection on which successful categorization is based. "Categorical perception," in which similarity space is "warped" in the service of categorization, turns out to be exhibited by both people and nets, and may mediate the constraints exerted by the analog world of objects on the formal world of symbols. [Retrieve by anonymous ftp in binary mode as (compressed) file harnad92.symbol.object.Z from directory /pub/harnad on princeton.edu] To retrieve a file by ftp from a Unix/Internet site, type: ftp princeton.edu When you are asked for your login, type: anonymous For your password, type your full name then change directories with: cd pub/harnad Then type: binary (This is for retrieving compressed files.) To show the available files, type: ls Next, retrieve the file you want with (for example): get filename.Z When you have the file(s) you want, type: quit Next uncompress the file with: uncompress filename.Z Now the file will be called, simply, filename --- The above cannot be done from Bitnet directly, but there is a fileserver called bitftp at pucc.bitnet that will do it for you. Send it the one line message help for instructions (which will be similar to the above, but will be in the form of a series of lines in an email message that bitftp will then execute for you). From sankar at bach.rutgers.edu Tue Apr 23 18:49:46 1991 From: sankar at bach.rutgers.edu (sankar@bach.rutgers.edu) Date: Tue, 23 Apr 91 18:49:46 EDT Subject: No subject Message-ID: <9104232249.AA05548@bach.rutgers.edu> The following two papers are now available via FTP from the neuroprose archives. Both will be presented at IJCNN-Seattle, 1991. These papers describe a new approach that combines Neural Networks and Decision Trees to form a classifier that grows the neurons as it learns. **************************************************************************** SPEAKER INDEPENDENT VOWEL RECOGNITION USING NEURAL TREE NETWORKS Ananth Sankar and Richard Mammone CAIP Center and Dept. of Electrical Engg. Rutgers University, P.O. Box 1390 Piscataway, NJ 08855-1390 Speaker independent vowel recognition is a difficult pattern recognition problem. Recently there has been much research using Multi-Layer Perceptrons (MLP) and Decision Trees for this task. This paper presents a new approach to this problem. A new neural architecture and learning algorithm called Neural Tree Networks (NTN) are developed. This network uses a tree structure with a neural network at each tree node. The NTN architecture offers a very efficient hardware implementation as compared to MLPs. The NTN algorithm grows the neurons while learning as opposed to backpropagation, for which the number of neurons must be known before learning can begin. The new algorithm is guaranteed to converge on the training set whereas backpropagation can get stuck in local minima. A gradient descent technique is used to grow the NTN. This approach is more efficient than the exhaustive search techniques used in standard decision tree algorithms. We present simulation results on a speaker independent vowel recognition task. These results show that the new method is superior to both MLP and decision tree methods. ***************************************************************************** OPTIMAL PRUNING OF NEURAL TREE NETWORKS FOR IMPROVED GENERALIZATION Ananth Sankar and Richard Mammone CAIP Center and Dept. of Electrical Engg. Rutgers University, P.O. Box 1390 Piscataway, NJ 08855-1390 An optimal pruning algorithm for a Neural Network recently developed called Neural Tree Networks (NTN) is presented. The NTN is grown by a constructive learning algorithm that decreases the classification error on the training data recursively. The optimal pruning algorithm is then used to improve generalization. The pruning algorithm is shown to be computationally inexpensive. Simulation results on a speaker independent vowel recognition task are presented to show the improved generalization using the pruning algorithm. *************************************************************************** To retrieve: unix> ftp cheops.cis.ohio-state.edu (or 128.146.8.62) Name: anonymous Password: neuron ftp> cd pub/neuroprose ftp> binary ftp> get sankar.ijcnn91_1.ps.Z ftp> get sankar.ijcnn91_2.ps.Z ftp> quit unix> uncompress sankar.ijcnn*.ps unix> lpr sankar.ijcnn91_1.ps sankar.ijcnn91_2.ps Thanks to Jordan Pollack for making this service available! From peterc at chaos.cs.brandeis.edu Tue Apr 23 22:58:55 1991 From: peterc at chaos.cs.brandeis.edu (Peter Cariani) Date: Tue, 23 Apr 91 22:58:55 edt Subject: THRESHOLDS AND SUPEREXCITABILITY In-Reply-To: Christof Koch's message of Thu, 18 Apr 91 13:59:13 PDT <910418135913.2040bf1d@Iago.Caltech.Edu> Message-ID: <9104240258.AA22066@chaos.cs.brandeis.edu> In reply to Christoff Koch's message: I had thought I had made myself clear regarding time-dependent threshold changes. Raymond and others have found a triphasic threshold time course following an action potential: the threshold initially goes up (refractory phase) far above its resting level, then it goes DOWN below its resting level (superexcitable phase), then it goes back up above its resting level (depression phase) and returns to its resting level. Superexcitability typically begins several milliseconds after the action potential and can last milliseconds to hundreds of milliseconds. While almost every neural model includes the refractory phase, I have yet to see one (outside of Raymond's model for threshold changes) which includes superexcitability. Actually, I had read Christoff Koch's and Tomasso Poggio's otherwise excellent review paper "Biophysics of computation: neurons, synapses and membranes." in Synaptic Function, Edelman, Wall & Cowan eds Wiley, 1989 when I began to look at neural models about a year and a half ago, and there is a section on conduction blocks but no mention of threshold changes or of any of the work by Lettvin & company. I know the book "Methods in Neuronal Modelling" and yes it has many detailed state-of-the-art models, but I explicitly looked there when the book first came out for a model incor- porating nonmonotonic threshold recovery curves and as I recall none of the models in that book even come close. This is not to say they are bad models (on the contrary, they are the best we have), but that there are some phenomena of potential import which are being overlooked. Peter Cariani From ken at cns.caltech.edu Wed Apr 24 19:00:04 1991 From: ken at cns.caltech.edu (ken@cns.caltech.edu) Date: Wed, 24 Apr 91 16:00:04 PDT Subject: 2nd announcement: Methods in Computational Neuroscience course Message-ID: <9104242300.AA16648@cns.caltech.edu> COURSE ANNOUNCEMENT: METHODS IN COMPUTATIONAL NEUROSCIENCE *** 2ND (AND LAST) NOTICE *** APPLICATIONS ARE DUE MAY 15 *** We are writing to bring to your attention the course ``Methods in Computational Neuroscience", to be offered Aug 5 -- 30, 1991 at the Marine Biological Laboratory, Woods Hole, Massachusetts. The course will provide intensive training in methods of neural modeling to 20 graduate students, postdocs, and faculty. Course directors are James Bower, Christof Koch, and Kenneth Miller (Associate Director), from Caltech. The course will include both lectures, and a laboratory in which each student will undertake a computational project of their own choosing. Each student will be provided with a Unix workstation (Sun Sparcstation II) and will be trained in doing neural simulations using the general purpose neural simulator GENESIS that has been developed at Caltech. GENESIS is publicly distributed without cost, so students can bring the skills acquired back to others at their institution. Course lectures are roughly divided into two sections. The first section will focus on biophysically realistic, detailed simulations of activity patterns in neurons and neuronal networks. Methods for simulating the electrical properties of single neurons will be introduced (compartmental models, active currents, interactions between synapses, calcium dynamics). Numerical and mathematical techniques necessary for such simulations will be presented, as will mathematical methods appropriate for understanding simplified network models (e.g. theory of dynamical systems). Applications to networks will be drawn from the invertebrate and vertebrate literature (central pattern generators, visual system of the fly, mammalian hippocampus, olfactory cortex). Faculty for this section will include: Paul Adams, SUNY Stony Brook Christof Koch, Caltech Idan Segev, Hebrew University Matthew Wilson, U. of Arizona Michael Mascagni, Supercomputing Research Center and NIH Jim Bower, Caltech Roger Traub, IBM John Rinzel, NIH Avis Cohen, U. Maryland Nancy Kopell, Boston U. Eve Marder, Brandeis Sylvia Ryckebusch, Caltech The second section will build on the first, focusing on learning, development, and higher functions. Topics such as LTP in hippocampus, development of the visual cortex and of the retino-tectal projection, and motion perception will be presented. Models at both the cellular and network levels will be discussed. More abstract networks, including biologically relevant learning algorithms and connectionist models, will also be examined from a neurobiological point of view. Faculty for this section will include: Charles Stevens, Salk Institute Anthony Zador, Yale Kenneth Miller, Caltech David MacKay, Caltech David Zipser, UCSD Terry Sejnowski, UCSD and the Salk Institute Christof Koch, Caltech Bill Bialek, NEC Corp. Rodney Douglas, Oxford David Van Essen, Caltech Richard Anderson, MIT Ted Adelson, MIT Requests for applications should be sent to: Ms. Florence Dwane Admissions Coordinator Marine Biological Laboratory Woods Hole, MA 02543 Applications should be sent to Ms. Dwane by May 15 1991. Financial aid is available for all students that need it. We believe this course provides a unique opportunity to develop skills in neural modeling. Please bring the course to the attention of your colleagues. We encourage you and your colleagues to consider attending. From lyle at ai.mit.edu Wed Apr 24 12:26:52 1991 From: lyle at ai.mit.edu (Lyle J. Borg-Graham) Date: Wed, 24 Apr 91 12:26:52 EDT Subject: THRESHOLDS AND SUPEREXCITABILITY In-Reply-To: Peter Cariani's message of Tue, 23 Apr 91 22:58:55 edt <9104240258.AA22066@chaos.cs.brandeis.edu> Message-ID: <9104241626.AA28990@wheat-chex> RE: Peter Cariani's message With the ever-increasing types (or as Paul Adams puts it, "Balkanization") of voltage/time/2nd-messenger(e.g. Ca++)-dependent channels being identified in various cells, it is clear that any canonical threshold model per se (tri-phasic time-dependency or otherwise) is problematic. The emerging picture is that there are a broad spectrum of channel combinations which may be participating in different contexts/neurons, resulting in quite different response modes. Whether a given model reproduces that spectrum or not depends, of course, on what the model is claiming to represent. The intricate and subtle interactions between channels in bullfrog slime neurons or hippocampal pyramidal cells (for example) are neat, and may *actually* be important for function. But that does not mean a much simpler membrane model would be more appropriate for a given simulation. Lyle Borg-Graham From koch at CitIago.Bitnet Thu Apr 25 05:51:29 1991 From: koch at CitIago.Bitnet (Christof Koch) Date: Thu, 25 Apr 91 02:51:29 PDT Subject: THRESHOLDS AND SUPEREXCITABILITY In-Reply-To: Your message <9104240258.AA22066@chaos.cs.brandeis.edu> dated 23-Apr-1991 Message-ID: <910425024432.2041a557@Iago.Caltech.Edu> I agree with Peter C. that none of the current models contain a superexcitable phase. However, do we know to what extent this phenomena occurs under physiological circumstances in say, cortical cells. As far as I remember Raymond and Lettvin's J. Physiol. paper on the timecourse of the spiking threshold, they worked on alpha motoneurons and made the cell fire hundreds of time. In a cortical cell, discharge of an action potential will transiently open calcium channels. The potential will return to the resting potential and will be pulled below to more hyperpolarizing levels by the combined action of several voltage and/or calcium-dependent potassium currents, in particular I_C, I_M and I_AHP. They effectively implement a relative refractory period. In hippocampal cells, action potentials are sometimes followed by ADP, i.e. after-depolarizations, caused by charge coming back from the dendrites. I wonder whether anybody knows whether the threshold for action potential initiation goes DOWN during these ADPs. This would be the closest phenomena to "superexcitability" I know of. Christof From lyle at ai.mit.edu Thu Apr 25 12:30:20 1991 From: lyle at ai.mit.edu (Lyle J. Borg-Graham) Date: Thu, 25 Apr 91 12:30:20 EDT Subject: THRESHOLDS AND SUPEREXCITABILITY In-Reply-To: Christof Koch's message of Thu, 25 Apr 91 02:51:29 PDT <910425024432.2041a557@Iago.Caltech.Edu> Message-ID: <9104251630.AA01291@substantia-nigra> > In a cortical cell, discharge of an action potential will transiently open > calcium channels. The potential will return to the resting potential > and will be pulled below to more hyperpolarizing levels by the combined > action of several voltage and/or calcium-dependent potassium currents, > in particular I_C, I_M and I_AHP. They effectively implement a relative > refractory period. > In hippocampal cells, action potentials are sometimes followed by ADP, i.e. > after-depolarizations, caused by charge coming back from the dendrites. Probably the reason a distinct "ADP" has been described in hippocampal pyramidal cells is that there is a ~1msec transient hyperpolarization (fAHP, or fast after-hyperpolarization) during the spike repolarization, due to the fast Ca-dependent K current (I_C). Since this current inactivates (possibly because of voltage) quickly, a subsequent depolarization results. However, the distributed charge of the dendrites which underlies this depolarization is most likely in other cells (the estimation of which is a good candidate for modelling), and thus the ADP phase is there as well. > I wonder whether anybody knows whether the threshold for action potential > initiation goes DOWN during these ADPs. I suppose it depends on what is meant by threshold, i.e. current or voltage. If the Na voltage 'threshold' doesn't change [which it does] then the ADP would lower the current threshold since, of course, the dendritic charge is lending a hand. > This would be the closest phenomena to "superexcitability" I know > of. For what it is worth, it might be better to reserve the term "excitability" for the state of intrinsic membrane properties. As far as "superexcitability" goes, perhaps the I_A and I_D K+ currents in hipppocampus may participate. The inactivation range of these currents include voltages below spike threshold, with varying time constants (I_A on the msec scale, I_D on the 10 to 100 msec scale). Thus spike activity tends to shut them off, which will contribute to excitability. Now the interesting thing to do is to identify the physiological pathways that modulate these guys. Lyle From SAYEGH at CVAX.IPFW.INDIANA.EDU Thu Apr 25 22:54:32 1991 From: SAYEGH at CVAX.IPFW.INDIANA.EDU (SAYEGH@CVAX.IPFW.INDIANA.EDU) Date: Thu, 25 Apr 1991 22:54:32 EDT Subject: proceedings 3rd NN & PDP + Air Mail Postage Message-ID: <910425225432.20202a97@CVAX.IPFW.INDIANA.EDU> In the announcement of the proceedings of the third conference on Neural Networks and Parallel Distributed Processing at Indiana-Purdue University, there was a minor mix-up in the list of papers. We also received a number of requests from Europe and Japan inquiring about Air Mail costs. Here is the info: Proceedings can be obtained by writing to: Ms. Sandra Fisher Physics Department Indiana University-Purdue University Ft Wayne, IN 46805 and including $5 + mailing and handling costs as follows $1 for the US, $2 for Canada and Mexico, $8.50 for all others by Air Mail or $3.95 by surface mail. Checks should be made payable to The Indiana-Purdue Foundation. The 109 page proceedings contain the following papers: INTEGRATED AUTONOMOUS NAVIGATION BY ADAPTIVE NEURAL NETWORKS Dean A. Pomerleau Department of Computer Science Carnegie Mellon University APPLYING A HOPFIELD-STYLE NETWORK TO DEGRADED PRINTED TEXT RESTORATION Arun Jagota Department of Computer Science State University of New York at Buffalo RECENT STUDIES WITH PARALLEL, SELF-ORGANIZ- ING, HIERARCHICAL NEURAL NETWORKS O.K. Ersoy & D. Hong School of Electrical Engineering Purdue University INEQUALITIES, PERCEPTRONS AND ROBOTIC PATH- PLANNING Samir I. Sayegh Department of Physics Indiana University-Purdue University GENETIC ALGORITHMS FOR FEATURE SELECTION FOR COUNTERPROPAGATION NETWORKS F.Z. Brill & W.N. Martin Department of Computer Science University of Virginia MULTI-SCALE VISION-BASED NAVIGATION ON DIS- TRIBUTED-MEMORY MIMD COMPUTERS A.W. Ho & G.C. Fox Caltech Concurrent Computation Program California Institute of Technology A NEURAL NETWORK WHICH ENABLES SPECIFICATION OF PRODUCTION RULES N. Liu & K.J. Cios The University of Toledo PIECE-WISE LINEAR ESTIMATION OF MECHANICAL PROPERTIES OF MATERIALS WITH NEURAL NETWORKS I.H. Shin, K.J. Cios, A. Vary* & H.E. Kautz* The University of Toledo & NASA Lewis Re- search Center* INFLUENCE OF THE COLUMN STRUCTURE ON INTRA- CORTICAL LONG RANGE INTERACTIONS E. Niebur & F. Worgotter California Institute of Technology LEARNING BY GRADIENT DESCENT IN FUNCTION SPACE Ganesh Mani University of Wisconsin-Madison SUCCESSIVE REFINEMENT OF MULTI-RESOLUTION REPRESENTATIONS OF THE ENVIRONMENT IN CONNECTIONIST NETWORKS Vasant Honovar and Leonard Uhr Computer Sciences Department University of Wisconsin-Madison A NEURAL ARCHITECTURE FOR COGNITIVE MAPS Martin Sonntag Cognitive Science & Machine Intelligence Lab University of Michigan From ken at cns.caltech.edu Fri Apr 26 10:20:41 1991 From: ken at cns.caltech.edu (ken@cns.caltech.edu) Date: Fri, 26 Apr 91 07:20:41 PDT Subject: Network: A New, Affordable Subscription Price Message-ID: <9104261420.AA18436@cns.caltech.edu> One of the several good neural network journals to have started in the last year or two is Network: Computation in Neural Systems, edited by Daniel Amit. The journal is self-consciously interdisciplinary between Neuroscience and physical/mathematical/computational sciences. The scope of the articles is similar to that of the articles found in Neural Computation, but Network includes primarily full papers (long articles) whereas Neural Computation includes primarily letters (short articles). My own interest in Network is motivated in no small part by the fact that the long, complete version of the paper of David MacKay and myself analyzing Linsker's simulations is found in Network, Vol. 1, #3; a short version was in Neural Computation. To mention a very few among many other recent authors: D. Willshaw, J. Hopfield, M. Abeles, P. Simic, D. Parisi, D. Amit, A. Selverston, G. Toulouse, V. Braitenberg, J. Nadal, E. Rolls. Network has been relatively inaccessible, at least in the US. I think this has been largely due to its very high subscription price. So I am writing to advertise the fact that Network has just instituted a very affordable subscription price for individuals. The new price for individuals is $37.90/year (US, Canada and Mexico), 17.30 lbs/year (UK), or 20.50 lbs/year (other). This price applies to Volume 1 (1990), as well as to the current volume, Vol. 2 (1991). To order: US/Can/Mex: American Institute of Physics, Subscriber Services, 500 Sunnyside Blvd, Woodbury, NY 11797-2999; 516-349-7800 x628; Other: Journals Mkting Dept., IOP Publishing, Techno House, Redcliffe Way, Bristol BS1 6NX, England. I won't spend bandwidth listing contents of recent issues; people at these addresses can send you literature. Ken Miller From ken at cns.caltech.edu Fri Apr 26 11:39:35 1991 From: ken at cns.caltech.edu (ken@cns.caltech.edu) Date: Fri, 26 Apr 91 08:39:35 PDT Subject: application material for Methods in Computational Neuroscience course Message-ID: <9104261539.AA18493@cns.caltech.edu> With respect to the previous 2nd announcement about the Methods of Computational Neuroscience course: I have received numerous inquiries as to how to obtain application materials rapidly (since the deadline for receipt of applications is soon -- May 15), so it seems merited to post the following info to the net as a whole: --- To repeat, the person who must be contacted to obtain application material, and to whom completed applications should be sent, is: Ms. Florence Dwayne, Admissions Coordinator, Marine Biological Laboratory, Woods Hole, MA 02543. [Note, I left the 'y' out of her last name in previous posting]. --- THERE IS NO E-MAIL ADDRESS FOR OBTAINING OR RETURNING APPLICATIONS. --- Ms. Dwayne can be reached by phone at 508-548-3705, extension 216. --- IF NECESSARY she can be reached by fax at 508-457-1924. The MBL fax machine is extremely busy, so they prefer not to send applications out by fax. However they will do so if necessary, particularly for overseas applicants. --- If completed applications must be returned quickly, please if possible use express mail or a courier service (i.e. Federal Express, DHL). Again, applications can IF NECESSARY be returned by fax, but this is strongly discouraged if an alternative is available. Ken Miller From worth at park.bu.edu Fri Apr 26 12:37:23 1991 From: worth at park.bu.edu (Andrew J. Worth) Date: Fri, 26 Apr 91 12:37:23 -0400 Subject: IJCNN-91-SEATTLE call for volunteers Message-ID: <9104261637.AA01878@park.bu.edu> --------------------------------------------------------------------------- IJCNN-91-Seattle Call for Volunteers July 8-12th, 1991 Seattle, Washington, USA. --------------------------------------------------------------------------- The International Joint Conference on Neural Networks (IJCNN-91-SEATTLE) has volunteer positions available. If you or anyone you know would like to exchange admittance to the conference for working as a volunteer, please respond directly to me at the e-mail address below. In the past, volunteers have given approximately 20 hours of labor (spread out over the entire conference) to receive: o admittance to the conference o a full set of proceedings o attendance to a limited number of tutorials (while working) The exact benefits are still being worked out. Volunteer positions include helping at: o Stuffing Conference Proceedings o Poster Sessions o Technical Sessions o Evening Plenary Sessions o Social Events o OPTIONAL duty: Tutorials If you are interested in volunteering, please respond directly to me with the following information: o Electronic Mail Address o Last Name, First Name o Address o Country o Phone number o Volunteer Position Preference Positions will be filled on a first commit first served basis. There will be no funding available for volunteer's travel and lodging expenses. PLEASE RESPOND TO: worth at park.bu.edu Thank you, Andy. ------------------------------------------------------------------------- Andrew J. Worth Cognitive & Neural Systems Dept. IJCNN-91 Volunteer Chair Boston University worth at park.bu.edu 111 Cummington Street, Rm 244 (617) 353-6741 Boston, MA 02215 ------------------------------------------------------------------------- From whsu at ecn.purdue.edu Sat Apr 27 12:02:25 1991 From: whsu at ecn.purdue.edu (William Hsu) Date: Sat, 27 Apr 91 11:02:25 -0500 Subject: Training standard one hideen layer backprop for Mglass Data Message-ID: <9104271602.AA10221@panther.ecn.purdue.edu> Dear Researchers, I have been trying to train a standard backprop network with one hidden layer with varying number of sigmoid units to learn the Mglass (Chaotic Data ref. Platt, Moody) Time series. I have not been able to acheive Normalized root mean square values of < 0.4 for archiectures to 50 hidden units. The architectures I have looked at includes 4-N-1. for various N. where the output unit is a linear summation unit (without sigmoid). I would appreciate hearing from people who have had similar experience or better still, people who have successful experience with the above setup. Pls email directly to me. Thank you very much. From booker at starbase.MITRE.ORG Mon Apr 29 14:51:13 1991 From: booker at starbase.MITRE.ORG (Lashon Booker) Date: Mon, 29 Apr 91 14:51:13 EDT Subject: ICGA-91 Preliminary Program / Call for Participation Message-ID: <9104291851.AA11805@starbase.mitre.org> Fourth International Conference on Genetic Algorithms ICGA91 13-16 July, 1991 Univ. California San Diego La Jolla, CA PRELIMINARY PROGRAM The Fourth International Conference on Genetic Algorithms (ICGA-91), will be held July 13-16, 1991 at the University of California - San Diego in La Jolla, CA. This meeting brings together an international community of scientists from academia and industry interested in algorithms suggested by the evolutionary process of natural selection. Topics of particular interest include: Holland's Genetic Algorithm and Classifier Systems, machine learning and optimization using these systems, and their relations to other learning paradigms (e.g., connectionist networks). Papers discussing how genetic algorithms and classifier systems are related to biological modeling issues (e.g., evolution of nervous systems, computational ethology, artificial life) have been encouraged. Interest in Genetic Algorithms (GAs) has risen steadily, and ICGA-91 drew more than 160 submissions. The conference committee believes it is especially important at this time of rapid growth in GA research to create a core of common understanding that will ensure future communication among the GA community. It is for this reason that ICGA-91 will mix invited talks and a small number of plenary presentations with as much small-group interaction as possible. Poster sessions will be the default mode for papers; some of the strongest papers to be included in ICGA-91 will be presented as posters. There will be no parallel sessions. CONFERENCE CO-CHAIRS: Kenneth A. De Jong J. David Schaffer Vice Chair: David E. Goldberg Program Co-Chairs: Richard K. Belew Lashon B. Booker Publicity Chair: David E. Goldberg Financial Chair: Gilbert Syswerda Local Arrangements: Richard K. Belew PROGRAM COMMITTEE: Emile Aarts, NV Phillips Richard K. Belew, UCSD Lashon B. Booker, MITRE Yuval Davidor, Weizmann Inst. L. David Davis, TICA Associates Kenneth A. De Jong, George Mason Univ. Larry Eschelman, Phillips Labs Stephanie Forrest, Univ. New Mexico John Greffenstette, Naval Research Lab. David E. Goldberg, Univ. Illinois Paulien Hogeweg, Univ. Utrecht John H. Holland, Univ. Michigan Gunar Liepins, Oakridge Natl. Lab. Heinz Muehlenbein, GMD, Germany John Koza, Stanford University Gregory Rawlins, Univ. Indiana Rick Riolo, Univ. Michigan George Robertson, Xerox PARC J. David Schaffer, Philips Labs Steve F. Smith, Carnegie-Mellon Univ. Gilbert Syswerda, BBN Tom Westerdale, Univ. London Stewart Wilson, Roland Inst. Darrell Whitley, Colorado State Univ. SUPPORTED BY: International Society for Genetic Algorithms Office of Naval Research Naval Research Laboratory TENTATIVE SCHEDULE Sat Sun Mon Tues AM 1: Tutorials Plenary 1 Plenary 3 Plenary 5 AM 2: Tutorials Poster 1 Poster 2 Poster 3 Lunch Lunch Lunch Lunch PM 1: Invited Talks Plenary 2 Plenary 4 Panel PM 2: Invited Talks Workshops Workshops Bus. Mtg. Evening: Social Mixer Workshops Banquet The schedule has been arranged so that it will be possible for participants to fly in Saturday (13 July) morning and leave Tuesday (16 July). We expect, however, that most ICGA-91 participants will want to attend the tutorials Saturday morning and the panel session and perhaps the ISGA business meeting Tuesday afternoon. Also, space is available for workshops after the business meeting and perhaps on Wednesday. Housing for these additional nights can be easily reserved at the same per-night rate; specify these on the registration form. INVITED SPEAKERS The ICGA-91 conference will officially begin on Saturday afternoon, 13 July with a series of invited presentations by leading researchers from computer science, mathematical biology, economics and philosophy: John Holland (University of Michigan) "Complex Adaptive Systems" Marcus Feldman (Stanford University) "Optimality and the Evolution of Recombination" John Miller (Santa Fe Institute) "Artificial Adaptive Agents in Economics" William Wimsatt (University of Chicago) "Developmental constraints on evolving systems" After these individual presentations and a dinner break, the guests will participate in a panel discussion delineating important research questions for the rest of the ICGA-91 meeting. TUTORIALS On Saturday morning, before the conference officially begins, we will be offering a series of tutorials. Two "tracks" have been designed, with one for ICGA-91 attendees who are new to Genetic Algorithms and one for attendees interested in advanced topics. An additional fee of $60 ($25 for students) will be charged to attend any and all tutorials; this fee includes course materials, lunch and the coffee break. Track I I.A Introduction to Genetic Algorithms (John Greffenstette) I.B Applications of Genetic Algorithms (L. David Davis) I.C Introduction to Classifier Systems (Rick Riolo) Track II II.A Genetic Algorithms Theory (Dave Goldberg and Gunar Liepins) II.B Biological Insights into GAs (Rik Belew and Chuck Talyor) PRE-REGISTRATION FOR THESE TUTORIALS (ON THE ENCLOSED REGISTRATION FORM) IS STRONGLY ADVISED, AS SPACE IS LIMITED. WORKSHOPS With the growing general interest in Genetic Algorithms (GAs), an important role for special interest groups is emerging. ICGA-91 has reserved time and space for smaller groups of people to meet and discuss some of these specialized aspects of GA research. Many of these topics were organized in response to large numbers of submissions in an area, many of which could not be published as part of the ICGA-91 Proceedings. If your work is related to one of these workshop topics, we encourage you to contact the chairperson of the relevant workshop. Also, if you have ideas for other workshops and would be willing to act as an organizer, contact Rik Belew (rik at cs.ucsd.edu) with your suggestion and we will do our best to publicize it. Here are the workshops presently organized: I. Machine Learning with GAs John Greffenstette (gref at aic.nrl.navy.mil) How well do GAs apply traditional machine learning problems, such as concept learning from examples, conceptual clustering, etc. How do GAs and classifier systems relate to other methods of reinforcement learning (e.g., Q-learning)? II. Evolutionary Strategie Hans-Paul Schwefel (schwefel at lumpi.informatik.uni-dortmund.de) A great deal of research, pioneered by Rechenberg, Schwefel and others, has been done in Germany on a class of "evolutionary strategies" closely related to Holland's form of the GA. This workshop is designed to acquaint members of the GA community unfamiliar with the theory and application of these techniques. III. Parallel Implementations of the GA Heinz Muhlenbein (muehlen at gmdzi.uucp) John McInerney (john at cs.ucsd.edu) GAs are proving themselves to be an important programming methodology for distributed and massively parallel machines. Moreover, this workshop will survey existing implementations of the GA on parallel machines and consider outstanding research issues. IV. GAs and Connectionist Networks Richard K. Belew (rik at cs.ucsd.edu) Both GAs and connectionist ("neural") networks represent active new approaches to machine learning. This workshop will consider hybrids of these methods, as well as attempts to relate GAs and connectionist networks as biological models of evolution and learning, resp. V. GAs and Artificial Life Peter Todd (todd at psych.stanford.edu) GAs are becoming central ingredients in a new class of "artificial life" simulations. This workshop will acquaint participants with some of the most important ALife experiments to date, and consider the special demands it makes on GAs. VI. Software support and test suites for GAs L. David Davis (70461.1552 at compuserve.com) Nici Schraudolf (nici at cs.ucsd.edu) Several large simulators have been developed for GAs, and a number of fitness function test suites have also been proposed. Both advances promise to make the job of the GA investigator easier, and also facilitate comparison across investigators. This workshop will survey these alternatives and attempt to form a consensus for future developments. VII. Biological influences in problem solving with Genetic Algorithms Alan Schultz (schultz at aic.nrl.navy.mil) The purpose of this workshop is to bring together GA researchers who are interested in improving the GA's performance in problem solving with mechanisms that are analogical to, or inspired by biological mechanisms. VIII. Theory of Genetic Operators Nick Radcliffe (njr at castle.edinburgh.ac.uk) There have been a number of recent developments which extend the applicability of schema analysis and our understanding of intrinsic parallelism. This workshop will examine these and look at new operators which have theoretical underpinning. TRAVEL INFORMATION Please make your travel arrangements early. Summer is a busy travel period in Southern California. San Diego can be reached conveniently by air, train or bus from Los Angeles. There are many direct flights to San Diego from around the United States. A special conference fare has been negotiated with American and Northwest Airlines for the ICGA conference. This applies to domestic and Canadian reservations only. A 40% discount will be offered for any roundtrip coach fare type, while 5% discount will be offered on the lowest available fare type, subject to capacity control, purchase time limits and other applicable fare rules. American Airlines: File Number 06718G; phone number: 1-800-433-1790. Northwest Airlines: File Number 04664; phone number: 1-800-328-1111. If you are arriving by air to the San Diego Airport (Lindbergh Field) look for a Peerless Shuttle; we have negotiated special rates for transportation to campus. Your registration confirmation package will include a flyer for your use. The rate is $14 for one person and $5 per additional person. A cab fare from the airport to UCSD is about $30. Ask the driver to take you to Warren College Conference Desk on Warren Campus. (You will probably need to show the campus map to the driver in order to find the registration desk, although he will know where UCSD is!) Housing registration can be completed at Warren College Conference Desk. Meeting registration will take place at La Casa during the Welcome Reception Friday evening or at the Price Center's Theatre Lobby on Saturday, Sunday or Monday. PARKING The distances on campus are not large and a car is not necessary. For example, the walk between the housing and the meeting facilities is about 3 minutes. However, if you are arriving by private or rental car, you will require a parking permit. These may be purchased at the Registration desk for $15.00 for a week or $4.00 per day. Parking permits are required seven days a week from 7am to 11pm. Monitoring of permits tends to be quite diligent. There is a parking moratorium on Saturday July 13 in parking lot 503 on Warren Campus; so, feel free to park anywhere there while you register. WEATHER The weather in La Jolla in July is usually very pleasant with temperatures around 25 degrees C (77 degrees F) and low humidity. Evenings can be cooler, so a light jacket or sweater may be useful. Occasionally very hot, dry weather (Santa Ana) can occur for a few days if the prevailing winds shift and come from the inland deserts. The Pacific Ocean is warm enough for swimming. We suggest you bring along swim-wear and a beach towel. REGISTRATION AND INFORMATION DESKS The registration and information desks will be open as follows: Conference registration will be in the Price Center Theatre Lobby: 7/12 - 7/15 7:00am - 5:00pm 7/16 7:00am - 2:00pm Housing registration only will be at the Warren College Desk, Daily 8:00am - 12:00 midnight If you need further assistance with some practical matter at any time, please see a member of the Organizing Committee. MESSAGES A telephone message service for incoming calls to conference participants will be set up. Please check the board in the foyer of the theater for messages. The conference telephone number will be communicated later. Outgoing calls may be made from public telephones located throughout the campus. If you have special problems with communication to the outside world during the conference, please stop by the registration desk or ask a member of the Organizing Committee. We are making arrangements to allow INTERNET access from terminals near the conference headquarters. REGISTRATION FEES The registration fee for conference participants includes entry to all technical sessions, conference proceedings, refreshment breaks, the receptions Friday and Saturday evenings, breakfast and lunch Sunday through Tuesday. The following fees have been established (all figures in U.S. dollars): Before After 6/1/91 6/1/91 Conference Participant $225 $275 Student $100 $125 A limited fund has been set aside to assist students and scientists with their travel expenses. If you are interested in obtaining such assistance and have no other form of travel support, send a letter describing your situation and needs to: Gilbert Syswerda International Society for Genetic Algorithms 53 Lake Street Winchester, MA 01890 (syswerda at bbn.com) Students should have their advisor certify their student status and that other funds are not available. REMITTANCE PROCEDURES Please fill out the registration form found elsewhere in this booklet and return with payment. Payment must be in U.S. dollars drawn on a U.S. bank. Checks should be made out to UC Regents, I.C.G.A. Personal checks, cashier's checks or international money orders in U.S. dollars are acceptable. Do not send cash. Cash (again in U.S. dollars) may be used for payment at the registration desk. Please note that credit cards cannot be accepted for payment under any circumstances. Do not send payment that is not in U.S. dollars. Registration form and payment should be sent to: Lene Hartman Conference Manager, I.C.G.A. UCSD 9500 Gilman Drive La Jolla, CA 92093-0513, USA The pre-registration deadline is June 1, 1991. Participants may register at a later time, including at the conference itself, although we strongly recommend that you register in advance. Mid-July still seems quite far away, but the program committee is already very excited by the prospect of a strong meeting of the GA community. Be sure to get your registration forms in early as attendance promises to be high and space is limited. Once the registration form and remittance have been received, the conference manager will send a receipt and directions to the participant's address. ON-CAMPUS HOUSING AND MEAL REGISTRATION INFORMATION Campus housing is available to conference participants as a package plan only - three nights' housing from Saturday through Monday. This housing is in two-bedroom apartments with shared living room, kitchen and bathroom. Each bedroom has two single beds and is rented as either a single (if only one person occupies the room) or as a double (if two people occupy the room). If you have made arrangements to share a double room with another I.C.G.A. participant, please include the name on the registration form. If you are willing to share an apartment but have not made such arrangements, we will coordinate such requests and assign you to a (same sex) apartment. Conference participants and accompanying guests may stay in campus housing several nights before or after the conference. Room rates for these additional nights are noted on the registration form. The rates given are per person, per night except for the family unit which is a per apartment, per night rate. Also make sure you specify your arrival and departure dates. A special lunch only package is available for accompanying guests. The cost of this package is $39 if purchased in advance; $64 if purchased after June 1. This purchase enables the accompanying guests to eat with the participants in the Price Center Ballroom. Accompanying guests choosing not to purchase meals as part of a package plan may purchase meals at one of the five restaurants within the Price Center or at one of the other Food Service facilities throughout campus. Additional tickets for the Receptions and the Banquet must be purchased in advance. BANQUET On Monday, 15 July a banquet will be held at Torrey Pines Inn. This hotel is perched overlooking the Pacific Ocean, and adjacent to the world famous Torrey Pines Golf course. The evening will begin with drinks around the Inn's Olympic swimming pool, watching the sun set into the Pacific. We will then be served a choice of beef, chicken or vegetarian dinners. (Please indicate your choice on the registration form.) Because we want to encourage attendance by the families of conference participants, we have made arrangements for a special children's room and dinner. Children will be served a choice of hamburger or vegetarian pizza, and supervised and entertained in a room of their own. BANQUET RESERVATIONS MUST BE MADE AS PART OF PRE-REGISTRATION. NO BANQUET TICKETS WILL BE AVAILABLE FOR SALE AT THE CONFERENCE. HOTEL ACCOMMODATIONS We have reserved a small number of rooms in the three hotels listed below: Torrey Pines Inn 11480 North Torrey Pines Road, La Jolla, CA 92037 Located next to the world-class Torrey Pines Golf Course and offers a quiet, rustic atmosphere. The distance between the Inn and UCSD is 1 1/2 miles and is a very pleasant walk. The Inn offers transportation to and from campus. The rate is $70 for double and $65 for single occupancy. Golf facilities and an olympic- size pool for lap swimming are available. Complimentary cocktail and newspaper. Please call (619) 453- 4420 for a reservation. La Jolla Village Inn, 3299 Holiday Court, La Jolla, CA 92037 Located 1/2 mile from UCSD this hotel offers complimentary transportation to and from the airport and the campus. The rate is $69 for double and $64 for single occupancy. Please call 1-800 854-2900 for a reservation. Sheraton Grande 10950 North Torrey Pines Rd., La Jolla, CA 92037 This new luxury hotel offers a fantastic view from each room of either the Pacific Ocean or the beautiful Torrey Pines Golf Course. Located atop Torrey Pines, it is convenient to UCSD campus and minutes from La Jolla. This exciting property is complete with several restaurants, cocktail lounges - all offering sunset views. Tennis, golf, swimming and a health spa are just a few of the hotel's amenities. The rate is $85 for either single or double. Please call (619) 558-1500 for a reservation. ON-CAMPUS RECREATION The campus housing is located near the recreation center which offers an olympic-size pool, racquetball and volleyball courts. Tennis courts are located near the housing complex (please bring your own racket and balls). Recreation cards may be purchased for a small fee ($14 single/$26 family) at the registration desk. PROGRAM FOR ACCOMPANYING GUESTS Transportation to Balboa Park, the San Diego Zoo, the Wild Animal Park and Old Town can be arranged. A modest fee will be charged. Sign-up sheets will be available at registration. The tours will be Sunday, Monday and Tuesday. Please sign up one day before your expected participation. GOLF The Torrey Pines Golf course is 1 1/2 miles from campus and the rates are: $35 per week-day, cash only $40 per week-end day $22 per cart $20 for clubs Reservations for a business day may be made two business days in advance at 1pm and for a week- end day five business days in advance. So, if you want to play Sunday, July 14 you must call Monday, July 8 at 8am. Call (619) 453-0380. INSURANCE Foreign participants are strongly encouraged to take out a traveller's health insurance policy before arriving in the United States. SMOKING UCSD has been designated by the university chancellor as a non-smoking environment. Smoking is not permitted within any public buildings on the campus. However, smoking is permitted in your apartment and on the grounds. SYNOPSIS OF IMPORTANT DATES May 1, 1991: Revised, camera-ready paper due June 1, 1991: Pre-registration deadline for conference and on-campus housing June 15, 1991: Audio visual and software demonstration requirements must be received July 13-16, 1991: Registration, conference starts ICGA91 - THE FOURTH INTERNATIONAL CONFERENCE ON GENETIC ALGORITHMS REGISTRATION FORM Dr./Mr./Ms.// Family Name:__________________ First Name: _________ Title: ___________________________ Institution: ____________________ Mailing Address: ___________________________________________________ Telephone: ______________________ FAX or Telex: ___________________ Electronic Mail Address: ___________________________________________ Name of Accompanying Person: _______________________________________ Preferred Roommate (for those choosing double rooms) _______________ Arrival Date: ____________________ Departure Date: _________________ ================================================================ CONFERENCE REGISTRATION Before On or After Quantity 6/1/91 6/1/91 Participant Registration ____ US $225 US $275 ________ Student Registration ____ US $100 US $125 ________ T-Shirt Size S ___ M ___ L ___ XL ___ Tutorial 7/13 Participant ____ US $ 60 US $ 85 ________ (Lunch included) Tutorial 7/13 Student ____ US $ 25 US $ 50 ________ (Lunch included) Track I ____ Track II ____ Registration sub-total ________ ================================================================ MEALS, RECEPTION, BANQUET Before On or After Quantity 6/1/91 6/1/91 Additional Welcome Reception Tickets Friday ____ US $10 US $10 ________ Additional Reception Tickets Saturday ____ US $10 US $10 ________ Lunch Only Package - 3 days (Accompanying Persons Only)____ US $39 US $64 ________ Banquet Ticket Adult ____ US $30 US $30 ________ Banquet Ticket Child (under 12) ____ US $ 8 US $ 8 ________ Meals sub-total ________ Banquet Choices: Adult: Beef ___ Chicken ___ Vegetarian ___ Child: Hamburger ___ Vegetarian ___ Pizza ___ PLEASE PURCHASE TICKET NOW IF YOU PLAN TO ATTEND. YOU WILL BE UNABLE TO BUY IT UPON ARRIVAL. ======================================================== APARTMENT HOUSING REGISTRATION Quantity Saturday through Monday (3 nights) Single Room $129/person ____ _________ Double Room $105/person ____ _________ Family(*) Unit $186/family ____ _________ Addl Single Nights $43/person/nite ____ _________ Addl Double Nights $35/person/nite ____ _________ Addl Family(*) Unit Nights $62/apt/nite ____ _________ (*) Family = 2 Adults and at least 1 child Housing sub-total _________ (Registrastion, Food & Housing) TOTAL _________ Payment by: U.S. Check ___ International Money Order ___ Cashier Check ___ Mail completed registration form and check made payable to "UC Regents" to: Lene Hartman, Conference Manager I.C.G.A. UCSD 9500 Gilman Drive La Jolla, CA 92093-0513 U.S.A. ======================================================== (for office use only) Check # _______ Date Rec'd _________ Balance Due _________ Initials ______ From schmidhu at informatik.tu-muenchen.dbp.de Tue Apr 30 03:17:00 1991 From: schmidhu at informatik.tu-muenchen.dbp.de (Juergen Schmidhuber) Date: 30 Apr 91 9:17 +0200 Subject: preprints and reports Message-ID: <9104300717.AA03329(a)kiss.informatik.tu-muenchen.de> Recent preprints and technical reports are available via ftp: ------------------------------------------------------------------ ADAPTIVE DECOMPOSITION OF TIME Juergen Schmidhuber, TUM (Talk at ICANN'91, Helsinki, June 24-28, 1991) In this paper we introduce design principles for unsupervised detection of regularities (like causal relationships) in temporal sequences. One basic idea is to train an adaptive predictor module to predict future events from past events, and to train an additional confidence module to model the reliability of the predictor's predictions. We select system states at those points in time where there are changes in prediction reliability, and use them recursively as inputs for higher-level predictors. This can be beneficial for `adaptive sub-goal generation' as well as for `conventional' goal-directed (supervised and reinforcement) learning: Systems based on these design principles were successfully tested on tasks where conventional training algorithms for recurrent nets fail. Finally we describe the principles of the first neural sequence `chunker' which collapses a self-organizing multi-level predictor hierarchy into a single recurrent network. LEARNING TO GENERATE SUBGOALS FOR ACTION SEQUENCES Juergen Schmidhuber, TUM (Talk at ICANN'91) This paper extends the technical report FKI-129-90 (`Toward compositional learning with neural networks'). USING ADAPTIVE SEQUENTIAL NEUROCONTROL FOR EFFICIENT LEARNING OF TRANSLATION AND ROTATION INVARIANCE Juergen Schmidhuber and Rudolf Huber, TUM (Talk at ICANN'91) This paper is based on FKI-128-90 (announced earlier). ------------------------------------------------------------------- LEARNING TO CONTROL FAST-WEIGHT MEMORIES: AN ALTERNATIVE TO DYNAMIC RECURRENT NETWORKS Juergen Schmidhuber, TUM Technical report FKI-147-91, March 26, 1991 Previous algorithms for supervised sequence learning are based on dynamic recurrent networks. This paper describes alternative gradient-based systems consisting of two feed-forward nets which learn to deal with temporal sequences by using fast weights: The first net learns to produce context dependent weight changes for the second net whose weights may vary very quickly. One advantage of the method over the more conventional recurrent net algorithms is the following: It does not necessarily occupy full-fledged units (experiencing some sort of feedback) for storing information over time. A simple weight may be sufficient for storing temporal information. Since with most networks there are many more weights than units, this property represents a potential for storage efficiency. Various learning methods are derived. Two experiments with unknown time delays illustrate the approach. One experiment shows how the system can be used for adaptive temporary variable binding. NEURAL SEQUENCE CHUNKERS Juergen Schmidhuber, TUM Technical report FKI-148-91, April 26, 1991 This paper addresses the problem of meaningful hierarchical adaptive decomposition of temporal sequences. This problem is relevant for time-series analysis as well as for goal-directed learning. The first neural systems for recursively chunking sequences are described. These systems are based on a principle called the `principle of history compression'. This principle essentially says: As long as a predictor is able to predict future environmental inputs from previous ones, no additional knowledge can be obtained by observing these inputs in reality. Only unpredicted inputs deserve attention. The focus is on a 2-network system which tries to collapse a self-organizing multi-level predictor hierarchy into a single recurrent network (the automatizer). The basic idea is to feed everything that was not expected by the automatizer into a `higher-level' recurrent net (the chunker). Since the expected things can be derived from the unexpected things by the automatizer, the chunker is fed with a reduced description of the input history. The chunker has a comparatively easy job in finding possibilities for additional reductions, since it works on a slower time scale and receives less inputs than the automatizer. Useful internal representations of the chunker in turn are taught to the automatizer. This leads to even more reduced input descriptions for the chunker, and so on. Experimentally it is shown that the system can be superior to conventional training algorithms for recurrent nets: It may require fewer computations per time step, and in addition it may require fewer training sequences. A possible extension for reinforcement learning and adaptive control is mentioned. An analogy is drawn between the behavior of the chunking system and the apparent behavior of humans. ADAPTIVE CONFIDENCE AND ADAPTIVE CURIOSITY Juergen Schmidhuber Technical Report FKI-149-91, April, 26, 1991 Much of the recent research on adaptive neuro-control and reinforcement learning focusses on systems with adaptive `world models'. Previous approaches, however, do not address the problem of modelling the reliability of the world model's predictions in uncertain environments. Furthermore, with previous approaches usually some ad-hoc method (like random search) is used to train the world model to predict future environmental inputs from previous inputs and control outputs of the system. This paper introduces ways for modelling the reliability of the outputs of adaptive world models, and it describes more sophisticated and sometimes much more efficient methods for their adaptive construction by on-line state space exploration: For instance, a 4-network reinforcement learning system is described which tries to maximize the future expectation of the temporal derivative of the adaptive assumed reliability of future predictions. The system is `curious' in the sense that it actively tries to provoke situations for which it {\em learned to expect to learn} something about the environment. In a very limited sense the system learns how to learn. An experiment with a simple non-deterministic environment demonstrates that the method can be clearly faster than the conventional model-building strategy. ------------------------------------------------------------------------ To obtain copies of the papers, do: unix> ftp 131.159.8.35 Name: anonymous Password: your name, please ftp> binary ftp> cd pub/fki ftp> get .ps.Z ftp> bye unix> uncompress .ps.Z unix> lpr .ps Here stands for any of the following six possibilities: icanndec (Adaptive Decomposition of Time) icannsub (Subgoal-Generator). This paper contains 5 partly hand-drawn figures which are not retrievable. Sorry. icanninv (Sequential Neuro-Control). fki147 (Fast Weights) fki148 (Sequence Chunkers) fki149 (Adaptive Curiosity) Please do not forget to leave your name. This will allow us to save paper if you are on our hardcopy mailing list. NOTE: icanninv.ps, fki148.ps, and fki149.ps are designed for European A4 paper format (20.9cm x 29.6cm). ------------------------------------------------------------------------ In case of ftp-problems contact Juergen Schmidhuber Institut fuer Informatik, Technische Universitaet Muenchen Arcisstr. 21 8000 Muenchen 2 GERMANY or send email to schmidhu at informatik.tu-muenchen.de DO NOT USE REPLY! From dfausett at zach.fit.edu Tue Apr 30 13:17:52 1991 From: dfausett at zach.fit.edu ( Donald W. Fausett) Date: Tue, 30 Apr 91 13:17:52 -0400 Subject: preprints and reports Message-ID: <9104301717.AA28357@zach.fit.edu> Greg: Thanks for the additional information. -- Don Fausett From well!moritz at apple.com Tue Apr 30 21:22:58 1991 From: well!moritz at apple.com (Elan Moritz) Date: Tue, 30 Apr 91 18:22:58 pdt Subject: Journal of Ideas, Vol 2 #1 Abstracts Message-ID: <9105010122.AA26275@well.sf.ca.us> +=++=++=++=++=++=++=++=++=++=++=++=++=++=++=+ please post & circulate Announcement ......... Abstracts of papers appearing in Volume 2 # 1 of the Journal of Ideas THOUGHT CONTAGION AS ABSTRACT EVOLUTION Aaron Lynch Abstract: Memory abstractions, or mnemons, form the basis of a memetic evolution theory where generalized self-replicating ideas give rise to thought contagion. A framework is presented for describing mnemon propagation, combination, and competition. It is observed that the transition from individual level considerations to population level considerations can act to cancel individual variations and may result in population behaviors. Equations for population memetics are presented for the case of two-idea interactions. It is argued that creativity via innovation of ideas is a population phenomena. Keywords: mnemon, meme, evolution, replication, idea, psychology, equation. ................... CULTURE AS A SEMANTIC FRACTAL: Sociobiology and Thick Description Charles J. Lumsden Department of Medicine, University of Toronto Toronto, Ontario, Canada M5S 1A8 Abstract: This report considers the problem of modeling culture as a thick symbolic system: a system of reference and association possessing multiple levels of meaning and interpretation. I suggest that thickness, in the sense intended by symbolic anthropologists like Geertz, can be treated mathematically by bringing together two lines of formal development, that of semantic networks, and that of fractal mathematics. The resulting semantic fractals offer many advantages for modeling human culture. The properties of semantic fractals as a class are described, and their role within sociobiology and symbolic anthropology considered. Provisional empirical evidence for the hypothesis of a semantic fractal organization for culture is discussed, together with the prospects for further testing of the fractal hypothesis. Keywords: culture, culturgen, meme, fractal, semantic network. ................... MODELING THE DISTRIBUTION OF A "MEME" IN A SIMPLE AGE DISTRIBUTION POPULATION: I. A KINETICS APPROACH AND SOME ALTERNATIVE MODELS Matthew Witten Center for High Performance Computing University of Texas System, Austin, TX 78758-4497 Abstract. Although there is a growing historical body of literature relating to the mathematical modeling of social and historical processes, little effort has been placed upon modeling the spread of an idea element "meme" in such a population. In this paper we review some of the literature and we then consider a simple kinetics approach, drawn from demography, to model the distribution of a hypothetical "meme" in a population consisting of three major age groups. KEYWORDS: Meme, idea, age-structure, compartment, sociobiology, kinetics model. ................... THE PRINCIPIA CYBERNETICA PROJECT Francis Heylighen, Cliff Joslyn, and Valentin Turchin The Principia Cybernetica Project[dagger] Abstract: This note describes an effort underway by a group of researchers to build a complete and consistent system of philosophy. The system will address, issues of general philosophical concern, including epistemology, metaphysics, and ethics, or the supreme human values. The aim of the project is to move towards conceptual unification of the relatively fragmented fields of Systems and Cybernetics through consensually-based philosophical development. Keywords: cybernetics, culture, evolution, system transition, networks, hypermedia, ethics, epistemology. ................... Brain and Mind: The Ultimate Grand Challenge Elan Moritz The Institute for Memetic Research P. O. Box 16327, Panama City, Florida 32406 Abstract: Questions about the nature of brain and mind are raised. It is argued that the fundamental understanding of the functions and operation of the brain and its relationship to mind must be regarded as the Ultimate Grand Challenge problem of science. National research initiatives such as the Decade of the Brain are discussed. Keywords: brain, mind, awareness, consciousness, computers, artificial intelligence, meme, evolution, mental health, virtual reality, cyberspace, supercomputers. +=++=++=++=++=++=++=++=++=++=++=++=++=++=++=+ The Journal of Ides an archival forum for discussion of 1) evolution and spread of ideas, 2) the creative process, and 3) biological and electronic implementations of idea/knowledge generation and processing. The Journal of Ideas, ISSN 1049-6335, is published quarterly by the Institute for Memetic Research, Inc. P. O. Box 16327, Panama City Florida 32406-1327. >----------- FOR MORE INFORMATION -------> E-mail requests to Elan Moritz, Editor, at moritz at well.sf.ca.us. +=++=++=++=++=++=++=++=++=++=++=++=++=++=++=+