From kmathia@gotham.accurate-automation.com Mon Mar 10 13:42:20 1997 Received: from lucy.cs.wisc.edu (lucy.cs.wisc.edu [128.105.2.11]) by sea.cs.wisc.edu (8.6.12/8.6.12) with ESMTP id NAA00927 for ; Mon, 10 Mar 1997 13:41:58 -0600 Received: from TELNET-1.SRV.CS.CMU.EDU (TELNET-1.SRV.CS.CMU.EDU [128.2.254.108]) by lucy.cs.wisc.edu (8.7.6/8.7.3) with SMTP id XAA06855 for ; Sun, 9 Mar 1997 23:39:53 -0600 (CST) Received: from TELNET-1.SRV.CS.CMU.EDU by telnet-1.srv.cs.CMU.EDU id aa27901; 9 Mar 97 21:53:54 EST Received: from DST.BOLTZ.CS.CMU.EDU by TELNET-1.SRV.CS.CMU.EDU id aa27892; 9 Mar 97 21:24:14 EST Received: from DST.BOLTZ.CS.CMU.EDU by DST.BOLTZ.CS.CMU.EDU id aa20688; 9 Mar 97 21:23:35 EST Received: from EDRC.CMU.EDU by B.GP.CS.CMU.EDU id aa22109; 7 Mar 97 17:09:24 EST Received: from gotham.accurate-automation.com by EDRC.CMU.EDU id aa13164; 7 Mar 97 17:08:53 EST Received: (from kmathia@localhost) by gotham.accurate-automation.com (8.7.5/8.7.3) id RAA07637; Fri, 7 Mar 1997 17:07:50 -0500 From: Karl Mathia Message-Id: <199703072207.RAA07637@gotham.accurate-automation.com> Subject: Learning Control at SCI'97 To: connectionists@cs.cmu.edu Date: Fri, 7 Mar 1997 17:07:50 -0500 (EST) Cc: corpora@hd.uib.no, cvnet@skivs.ski.org, cybsys-l@bingvmb.cc.binghamton.edu, empiricists@csli.stanford.edu, eletter-request@win.tue.nl, hybrid-list@cs.ua.edu, inns-l%umdd.bitnet@pucc.princeton.edu, issnnet-mlist@CNS.BU.EDU, ml@ics.uci.edu X-Mailer: ELM [version 2.4 PL25] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit ! PLEASE POST ! PLEASE POST ! PLEASE POST ! PLEASE POST ! PLEASE POST ! *********************************************************************** LAST Call for Papers: SPECIAL SESSION ON LEARNING CONTROL ----------------------------------- at the WORLD MULTICONFERENCE ON SYSTEMICS, CYBERNETICS AND INFORMATICS (SCI'97) Caracas, Venezuela July 7-11, 1997 (Extended Deadline for Abstracts: March 31, 1997) *********************************************************************** LEARNING CONTROL ---------------- "Learning Control" is a term attributed to a broad class of self-tuning processes, where the performance of the controlled system with respect to a particular task is self-improved based on the performance for previous identical tasks. The idea of self-learning control systems is aesthetically appealing and represents a fundamental step towards fully autonomous systems. This is of advantage when dealing with uncertain or changing systems. The major difference between adaptive and learning control is sometimes characterized in terms of 'local' and 'global' learning. An adaptive systems continuously adapts to changes in environment and system para- meters (local), whereas a learning systems memorizes and recognizes previously experienced situations (global). The classification of learning control will be one of many topics at this SCI'97 session. We invite you to present your recent research results, learn about current avenues in the field, and to meet interesting people. We look for quality papers which cover the topics outlined below. Similar work is also welcome. OVERVIEW OF SCI'97 ------------------ SCI'97 is a truly multi-disciplinary conference, covering intelligent computing, information theory, cybernetics, social and biological systems, psychology, and applications. General information about the conference is listed below or can be found at the website http://www.iiis.org/ SCI'97 is an ideal platform for a special session on "Learning Control", an emerging discipline which is receiving more and more attention from both the academic and industrial controls community, due to the increasing complexity of (technical) systems. TOPICS ------ The Learning Control Session will include, but is not limited, to the following topics (further suggestions are encouraged): * Classification of learning control systems. * Mathematical learning theory in a controls context. * Biological or social self-learning control mechanisms and their extension to technical systems. * Human operator modeling. Human operators are (currently) the ultimate learning controller for complex systems. * Neurocontrol, using biological or artificial neural networks. * Fuzzy logic and learning. * Optimal Control type learning algorithms (adaptive critics, Q-learning, etc.). * Variable structure learning of controllers and its variants, e.g. reconfigurable and reparameterizable controllers. * Stability of learning control systems (important!). * Hardware implementations. * Applications and case studies which exceed the usual benchmark problems towards real-world complex systems. Questions about, or contributions to this special session can be e-mailed to Karl Mathia at karl@mathia.com or kmathia@accurate-automation.com PAPER SUBMISSION ---------------- Please mail three (3) hardcopies of your abstract or draft (1-2 pages) to: Dr. Karl Mathia Accurate Automation Corporation 7001 Shallowford Road Phone: (423) 894-4646 Chattanooga, TN 37421 Fax: (423) 894-4645 USA Full-size papers (max. 8 pages, single spaced) are be submitted by authors after the notification of acceptance. Please note the extended deadline for camera-ready papers: May 12, 1997. DEADLINES --------- March 15, 1997 Submission of 1-2 page abstracts or drafts. March 31, 1997 Acceptance notifications. May 1, 1997 Submission of camera ready papers (max. 8 pages, single spaced). *********************************************************************** From jordan@psyche.mit.edu Mon Mar 10 13:44:24 1997 Received: from lucy.cs.wisc.edu (lucy.cs.wisc.edu [128.105.2.11]) by sea.cs.wisc.edu (8.6.12/8.6.12) with ESMTP id NAA01029 for ; Mon, 10 Mar 1997 13:43:56 -0600 Received: from TELNET-1.SRV.CS.CMU.EDU (TELNET-1.SRV.CS.CMU.EDU [128.2.254.108]) by lucy.cs.wisc.edu (8.7.6/8.7.3) with SMTP id WAA05543 for ; Sun, 9 Mar 1997 22:22:23 -0600 (CST) Received: from TELNET-1.SRV.CS.CMU.EDU by telnet-1.srv.cs.CMU.EDU id aa27890; 9 Mar 97 21:45:04 EST Received: from DST.BOLTZ.CS.CMU.EDU by TELNET-1.SRV.CS.CMU.EDU id aa27888; 9 Mar 97 21:23:14 EST Received: from DST.BOLTZ.CS.CMU.EDU by DST.BOLTZ.CS.CMU.EDU id aa20680; 9 Mar 97 21:22:38 EST Received: from RI.CMU.EDU by B.GP.CS.CMU.EDU id aa02919; 8 Mar 97 10:13:14 EST Received: from PSYCHE.MIT.EDU by RI.CMU.EDU id aa24342; 8 Mar 97 10:12:28 EST Received: by psyche.mit.edu (4.1/DA1.0.4) id AA29774; Sat, 8 Mar 97 10:12:25 EST From: Michael Jordan Message-Id: <9703081512.AA29774@psyche.mit.edu> Subject: NIPS*97 Call for Papers To: connectionists@cs.cmu.edu Date: Sat, 8 Mar 97 10:12:24 EST Cc: Michael Jordan X-Mailer: ELM [version 2.3 PL0] CALL FOR PAPERS -- NIPS*97 Neural Information Processing Systems -- Natural and Synthetic Monday December 1 - Saturday December 6, 1997 Denver, Colorado This is the eleventh meeting of an interdisciplinary conference which brings together cognitive scientists, computer scientists, engineers, neuroscientists, physicists, and mathematicians interested in all aspects of neural processing and computation. The conference will include invited talks and oral and poster presentations of refereed papers. The conference is single track and is highly selective. Preceding the main session, there will be one day of tutorial presentations (Dec. 1), and following will be two days of focused workshops on topical issues at a nearby ski area (Dec. 5-6). Major categories for paper submission, with example subcategories (by no means exhaustive), are as follows: Algorithms and Architectures: supervised and unsupervised learning algorithms, model selection algorithms, feedforward and recurrent network architectures, localized basis functions, online learning algorithms, active learning algorithms, algorithms for combining classifiers, belief networks, combinatorial optimization. Applications: handwriting recognition, DNA and protein sequence analysis, expert systems, fault diagnosis, financial analysis, medical diagnosis, music processing, time-series prediction. Artificial Intelligence: inductive reasoning, problem solving and planning, natural language understanding, hybrid symbolic-subsymbolic systems. Cognitive Science: perception and psychophysics, development, neuropsychology, cognitive neuroscience, language, human learning and memory, attention. Implementation: analog and digital VLSI, optical neurocomputing systems, novel neuro-devices, simulation tools, parallelism. Neuroscience: functional imaging, systems physiology, neural coding, synchrony, synaptic plasticity, neuromodulation, dendritic computation, calcium dynamics, inhibition, computational models. Reinforcement Learning and Control: exploration, dynamic programming, planning, navigation, robotic motor control, process control, Markov decision processes. Speech and Signal Processing: speech recognition, speech coding, speech synthesis, rapid adaptation, robust processing, auditory scene analysis, models of human speech perception. Theory: computational learning theory, statistical mechanics of learning, dynamics of learning algorithms, learning of dynamical systems, approximation and estimation theory, combining predictors, model selection, complexity theory. Visual Processing: image processing, image coding and classification, object recognition, stereopsis, motion detection and tracking, visual psychophysics. Review Criteria: All submitted papers will be thoroughly refereed on the basis of technical quality, significance, and clarity. Novelty of the work is also a strong consideration in paper selection, but to encourage interdisciplinary contributions, we will consider work which has been submitted or presented in part elsewhere, if it is unlikely to have been seen by the NIPS audience. Authors should not be dissuaded from submitting recent work, as there will be an opportunity after the meeting to revise accepted manuscripts before submitting final camera-ready copy. Paper Format: Submitted papers may be up to seven pages in length, including figures and references, using a font no smaller than 10 point. Submissions failing to follow these guidelines will not be considered. Authors are encouraged to use the NIPS LaTeX style files obtainable by anonymous FTP at the site given below. Papers must indicate (1) physical and e-mail addresses of all authors; (2) one of the nine major categories listed above, and, if desired, a subcategory; (3) if the work, or any substantial part thereof, has been submitted to or has appeared in other scientific conferences; (4) the authors' preference, if any, for oral or poster presentation (this preference will play no role in paper acceptance); and (5) author to whom correspondence should be addressed. Submission Instructions: Send eight copies of submitted papers to the address below; electronic or FAX submission is not acceptable. Include one additional copy of the abstract only, to be used for preparation of the abstracts booklet distributed at the meeting. SUBMISSIONS MUST BE RECEIVED BY MAY 23, 1997. From within the U.S., submissions will be accepted if mailed first class and postmarked by May 20, 1997. Mail submissions to: Michael Kearns NIPS*97 Program Chair AT&T Laboratories Research Room 2A-423 600 Mountain Avenue Murray Hill, NJ 07974-0636 USA Mail general inquiries and requests for registration material to: NIPS*97 Registration Conference Consulting Associates 451 N. Sycamore Monticello, IA 52310 fax: (319) 465-6709 (attn: Denise Prull) e-mail: nipsinfo@salk.edu Copies of the LaTeX style files for NIPS are available via anonymous ftp at ftp.cs.cmu.edu (128.2.206.173) in /afs/cs/Web/Groups/NIPS/formatting The style files and other conference information may also be retrieved via World Wide Web at http://www.cs.cmu.edu/Web/Groups/NIPS NIPS*97 Organizing Committee: General Chair, Michael Jordan, MIT; Program Chair, Michael Kearns, AT&T Labs Research; Publications Chair, Sara Solla, Northwestern University; Tutorial Chair, Satinder Singh, University of Colorado; Workshops Co-Chairs, Steven Nowlan, Lexicus, and Richard Zemel, University of Arizona; Publicity Chair, Anthony Bell, Salk Institute; Local Arrangements, Arun Jagota, University of California, Santa Cruz; Treasurer, Bartlett Mel, University of Southern California; Web Master, Doug Baker, Carnegie Mellon University; Government Liaison, John Moody, OGI; Contracts, Steve Hanson, Rutgers University, Scott Kirkpatrick, IBM, Gerry Tesauro, IBM. Conference arrangements by Conference Consulting Associates, Monticello, IA. NIPS*97 Program Committee: Sue Becker, McMaster University; Joachim Buhmann, University of Bonn; Tom Dietterich, Oregon State University; Michael Kearns, AT&T Labs Research (chair); Richard Lippmann, MIT Lincoln Lab; Larry Saul, AT&T Labs Research; Jude Shavlik, University of Wisconsin; Rich Sutton, University of Massachusetts; Tali Tishby, Hebrew University; Michael Turmon, Jet Propulsion Lab; Paul Viola, MIT; John Wawrzynek, UC Berkeley; Tony Zador, Salk Institute. DEADLINE FOR RECEIPT OF SUBMISSIONS IS MAY 23, 1997 - please post - From paolo@eealab.unian.it Mon Mar 10 15:25:39 1997 Received: from lucy.cs.wisc.edu (lucy.cs.wisc.edu [128.105.2.11]) by sea.cs.wisc.edu (8.6.12/8.6.12) with ESMTP id PAA00363 for ; Mon, 10 Mar 1997 15:25:10 -0600 Received: from TELNET-1.SRV.CS.CMU.EDU (TELNET-1.SRV.CS.CMU.EDU [128.2.254.108]) by lucy.cs.wisc.edu (8.7.6/8.7.3) with SMTP id PAA21760 for ; Mon, 10 Mar 1997 15:25:09 -0600 (CST) Received: from TELNET-1.SRV.CS.CMU.EDU by telnet-1.srv.cs.CMU.EDU id aa29134; 10 Mar 97 13:47:43 EST Received: from DST.BOLTZ.CS.CMU.EDU by TELNET-1.SRV.CS.CMU.EDU id aa29132; 10 Mar 97 13:34:24 EST Received: from DST.BOLTZ.CS.CMU.EDU by DST.BOLTZ.CS.CMU.EDU id aa21308; 10 Mar 97 13:34:21 EST Received: from RI.CMU.EDU by B.GP.CS.CMU.EDU id aa03667; 10 Mar 97 12:44:56 EST Received: from icineca.cineca.it by RI.CMU.EDU id aa04333; 10 Mar 97 12:43:45 EST Received: from dns.eealab.unian.it by ICINECA.CINECA.IT (IBM VM SMTP V2R2) with TCP; Mon, 10 Mar 97 18:39:59 ITA Received: by dns.eealab.unian.it (4.1/SMI-4.1) id AA00234; Mon, 10 Mar 97 17:07:22 +0200 Date: Mon, 10 Mar 97 17:07:22 +0200 From: Paolo Campolucci Message-Id: <9703101507.AA00234@dns.eealab.unian.it> Apparently-To: connectionists@cs.cmu.edu --------------------------CALL FOR PAPERS-------------------------------- Special Session on Neural Networks for Signal Processing ---------------------------------------------------------------------------- --------- ISIS'97- International Symposium on Intelligent Systems ---------------------------------------------------------------------------- --------- Signal processing problems, with their particular features and challenges, have provided an important field of application for artificial neural networks (ANNs). The non-linear processing and classification capabilities of these networks, in fact, can be very useful in many DSP applications. However, to provide robust, efficient and reliable NN solutions to many real-world problems, new and original ideas should be found to overcome nowadays limitations. This special session, organized within ISIS'97, wants to provide an open forum for researchers in this field to present and discuss new and promising results on the theory and application of artificial neural networks for signal processing. Papers are solicited for, but not limited to, the following topics: >>>Paradigms: artificial neural networks, evolutionary computation, nonlinear signal processing, robust fitting, classification and learning, statistical properties of ANNs, complexity issues. >>>Application areas: speech processing, speech enhancement, space-time processing, adaptive filtering, channel equalization and predistortion, robotics, system identification and control, time series prediction, temporal pattern recognition. >>>Theories: generalization, design algorithms, optimization, learning, neural architectures for signal processing, dynamic recurrent neural networks, locally recurrent neural networks, IIR synapses neural networks, Time Delay Neural Networks, fast learning algorithms, on-line training algorithms, on-line clustering, and identification/rejection of outliers. >>>Implementations: parallel and distributed implementation, hardware design, other implementation technologies. The International Symposium on Intelligent Systems will be held in Reggio Calabria (Italy) in the center of the Mediterranean sea, from September 11 to September 13, 1997. This 2+1/2 days interdisciplinary symposium aims to be an international forum for advanced studies in Neural Networks, Fuzzy Systems and related intelligent technologies, with special concern for analysis and synthesis of systems which have to make decisions in an environment of uncertainty and imprecision. This Conference follows up the series of International Symposia organized by AMSE in Istanbul (1988), Brighton (1989), Cetinje (1990), Warsaw (1991), Geneva (1992), London (1993), Lyon (1994), Brno (1995) and Leon (1996). More information about the workshop is available at: http://neurolab.ing.unirc.it/isis97.html Papers Submission to special session: Papers must be received by April 15, 1997. They should include an abstract not exceeding 100 words and should not exceed five A4 pages, single-column format in Times Roman or similar font style, 10 points or larger with 2.5 cm (one inch) margins on all four sides, including figures, tables and references. All submitted manuscripts, both invited and contributed, will be evaluated by peer reviewers to determine their suitability for publication. Authors are encouraged to submit their work via Air Mail or Express Courier so as to ensure timely delivery. All submissions will be acknowledged by electronic or postal mail. Authors of accepted papers wil= be invited to submit their final camera-ready papers by May 31, 1997. Five copies of the manuscript must be sent to the address below with an accompanying letter including the following information: Full Title of the Paper Technical Topics (First and Second Choices) Corresponding Author (Name, Postal and E-Mail Addresses, Telephone and Fax Numbers) Presenting Author and preferred mode of presentation (Oral or Poster) Symposium Language: English will be the official language of the Symposium. Important dates: April 15, 1997 Deadline for manuscripts submitted to the special session. May 10, 1997 Notification to Authors.=20 May 31, 1997 Deadline for receiving final camera-ready papers Proceedings: Accepted papers will be included in the Proceedings published by IOS press. Selected papers could be published in several journals of AMSE. All manuscripts submitted to the special session must be sent to: Prof. Francesco Piazza Dipart. Elettronica ed Automatica, Universita' di Ancona via Brecce Bianche, 60131 Ancona email: upf@eealab.unian.it (or paolo@eealab.unian.it) TEL: +39 (71) 2204 453 or 2204 541 FAX: +39 (71) 2204 464 or 2204 835 For informations about the special session, please contact Prof. F. Piazza, at the address above, or: Dr. Elio D. Di Claudio Dipart. INFO-COM, Universit=E0 di Roma "La Sapienza" via Eudossiana 18, 00184 Roma email dic@infocom.ing.uniroma1.it TEL: +39 (6) 44585 837 or 44585 839 FAX: +39 (6) 4873300 For information about the symposium, please contact: Symposium Secretariat: ISIS'97 Secretariat=20 University of Reggio Calabria=20 Faculty of Engineering, DIEMA=20 Via E.Cuzzocrea 48, I-89127 Reggio Calabria, Italy Phone: +39 -965 875224, Fax: +39 -965 875247 E-mail: neurolab@csiins.unirc.it WWW: http://neurolab.ing.unirc.it/isis97.html From henkel@physik.uni-bremen.de Tue Mar 11 09:31:56 1997 Received: from lucy.cs.wisc.edu (lucy.cs.wisc.edu [128.105.2.11]) by sea.cs.wisc.edu (8.6.12/8.6.12) with ESMTP id JAA00712 for ; Tue, 11 Mar 1997 09:31:31 -0600 Received: from TELNET-1.SRV.CS.CMU.EDU (TELNET-1.SRV.CS.CMU.EDU [128.2.254.108]) by lucy.cs.wisc.edu (8.7.6/8.7.3) with SMTP id HAA09325 for ; Tue, 11 Mar 1997 07:30:45 -0600 (CST) Received: from TELNET-1.SRV.CS.CMU.EDU by telnet-1.srv.cs.CMU.EDU id aa00367; 11 Mar 97 5:48:22 EST Received: from DST.BOLTZ.CS.CMU.EDU by TELNET-1.SRV.CS.CMU.EDU id aa00365; 11 Mar 97 5:27:44 EST Received: from DST.BOLTZ.CS.CMU.EDU by DST.BOLTZ.CS.CMU.EDU id aa22164; 11 Mar 97 5:26:08 EST Received: from EDRC.CMU.EDU by B.GP.CS.CMU.EDU id aa18730; 11 Mar 97 4:47:36 EST Received: from theo.physik.uni-bremen.de by EDRC.CMU.EDU id aa01290; 11 Mar 97 4:47:13 EST Received: from axon.physik.uni-bremen.de by theo.physik.uni-bremen.de (AIX 3.2/UCB 5.64/4.03vk7) id AA25667; Tue, 11 Mar 1997 10:43:42 +0100 Sender: rdh@theo.physik.uni-bremen.de Message-Id: <332537CA.12A4AAE6@theo.physik.uni-bremen.de> Date: Tue, 11 Mar 1997 11:45:30 +0100 From: Rolf Henkel Organization: Institute for Theoretical Neurophysics X-Mailer: Mozilla 3.0 (X11; I; Linux 2.0.29 i586) Mime-Version: 1.0 To: connectionists@cs.cmu.edu Subject: Stereovision: TR, Webpages, Demo Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit A technical report, some Webpages and an Online-Demo of a new computational approach to Stereovision are now available. The new approach rests on aliasing effects of simple disparity estimators caused by sampling visual space only at two eye-positions. In connection with a coherence-detection scheme, a stable algorithm is obtained which calculates within a single computational structure a dense disparity map and a verification count for the disparity estimates. In addition, the network fuses the left and right stereo images into the cyclopean view of the scene. Keywords: Stereovision, cyclopean view, complex cells, parallel algorithm Comments to the ideas presented are very welcome. The webpages can be found at the URL http://axon.physik.uni-bremen.de/~rdh/research/stereo/ the technical report retrieved under http://axon.physik.uni-bremen.de/~rdh/research/papers/tyc.ps.gz and the Online-Demo of the algorithm tested under the URL http://axon.physik.uni-bremen.de/~rdh/online_calc/stereo2/ Thank you very much for your interest, Rolf Henkel Institute of Theoretical Neurophysics, University Bremen, Germany ----------------------------------------------------------------- Email: henkel@theo.physik.uni-bremen.de URL: http://axon.physik.uni-bremen.de/~rdh/research/ ----------------------------------------------------------------- From golden@utdallas.edu Tue Mar 11 18:11:02 1997 Received: from lucy.cs.wisc.edu (lucy.cs.wisc.edu [128.105.2.11]) by sea.cs.wisc.edu (8.6.12/8.6.12) with ESMTP id SAA17667 for ; Tue, 11 Mar 1997 18:10:56 -0600 Received: from TELNET-1.SRV.CS.CMU.EDU (TELNET-1.SRV.CS.CMU.EDU [128.2.254.108]) by lucy.cs.wisc.edu (8.7.6/8.7.3) with SMTP id SAA25881 for ; Tue, 11 Mar 1997 18:10:54 -0600 (CST) Received: from TELNET-1.SRV.CS.CMU.EDU by telnet-1.srv.cs.CMU.EDU id aa01131; 11 Mar 97 18:02:47 EST Received: from DST.BOLTZ.CS.CMU.EDU by TELNET-1.SRV.CS.CMU.EDU id aa01129; 11 Mar 97 17:48:48 EST Received: from DST.BOLTZ.CS.CMU.EDU by DST.BOLTZ.CS.CMU.EDU id aa22515; 11 Mar 97 17:47:40 EST Received: from EDRC.CMU.EDU by B.GP.CS.CMU.EDU id ab28694; 11 Mar 97 13:49:30 EST Received: from utdallas.edu by EDRC.CMU.EDU id aa04137; 11 Mar 97 13:48:49 EST Received: (from root@localhost) by utdallas.edu (8.8.5/8.8.5) id MAA19122; Tue, 11 Mar 1997 12:48:33 -0600 (CST) Received: from medulla.utdallas.edu by utdallas.edu (Brelay v6.01) with BLIMP; Tue, 11 Mar 1997 12:48:32 CST Date: Tue, 11 Mar 1997 12:48:32 -0600 (CST) From: Richard M Golden To: connect-bb@eusip.ed.ac.uk, connectionists@cs.cmu.edu, crtnet%psuvm.BITNET@pucc.princeton.edu, cti-econ@mailbase.ac.uk, cybsys-l%bingvmb.BITNET@pucc.princeton.edu Subject: NEW MIT BOOK! Mathematical Methods for Neural Net Analysis and Design Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Mathematical Methods for Neural Network Analysis and Design Richard M. Golden MIT Press (1996) ORDERING INFO: ISBN 0-262-07174-6 http://www.amazon.com Cloth ($65.00), 432 pages. http://www-mitpress.mit.edu/order-info.html 1-800-356-0343 (MIT BOOK ORDER Department) ***FOR MORE INFO: http://www.utdallas.edu/~golden/book_abs.html This textbook teaches students how to carefully use a powerful set of mathematical tools for analyzing and designing a wide variety of NONLINEAR HIGH-DIMENSIONAL Artificial Neural Network (ANN) systems. Chapter 1: ANN systems with Neuroscience, Psychology, Engineering Applications Chapter 2: Specific ANN system architectures; classification/learning paradigms Chapter 3: LaSalle's Invariant Set Theorem for behavioral analysis of ANNs Chapter 4: Stochastic Approximation Theorem for behavioral analysis of ANNs Chapter 5: Nonlinear Optimization Theory for ANN system design Chapters 6,7: Bayesian Decision Theory and Markov Random Fields for the design of "rational" ANN classification/learning objective functions. Chapter 8: Confidence intervals for an ANN system's predictions. Statistical tests for: (i) pruning/adding units, and (ii) model selection. Solutions: Solutions to over 100 ANN system analysis and design problems ***** This message is being sent to multiple mailing lists. My apologies if you receive this message more than once. ****** From Ilpnet@ijs.si Wed Mar 12 12:24:06 1997 Received: from lucy.cs.wisc.edu (lucy.cs.wisc.edu [128.105.2.11]) by sea.cs.wisc.edu (8.6.12/8.6.12) with ESMTP id MAA06478 for ; Wed, 12 Mar 1997 12:23:59 -0600 Received: from uhura.ijs.si (uhura.ijs.si [193.2.4.252]) by lucy.cs.wisc.edu (8.7.6/8.7.3) with SMTP id MAA14398 for ; Wed, 12 Mar 1997 12:23:57 -0600 (CST) Received: from cathy.ijs.si by uhura.ijs.si with SMTP using DNS (PP) id <02342-0@uhura.ijs.si>; Wed, 12 Mar 1997 19:22:09 +0100 Received: from CATHY.IJS.SI by CATHY.IJS.SI (PMDF V4.3-10 #8779) id <01IGFCRSUS2O005I20@CATHY.IJS.SI>; Wed, 12 Mar 1997 19:22:04 +0001 Date: Wed, 12 Mar 1997 19:22:04 +0001 From: "ILPNET project, IJS" Subject: ILPNewsletter Vol. 4 No. 1 To: ml@cs.wisc.edu Message-id: <01IGFCRSWDXU005I20@CATHY.IJS.SI> X-VMS-To: in%"ml@cs.wisc.edu" MIME-version: 1.0 Content-type: TEXT/PLAIN; CHARSET=US-ASCII Content-transfer-encoding: 7BIT %------------------------------------------------------------------------------% ILP Newsletter Volume 4, Number 1, 5th March 1997 %------------------------------------------------------------------------------% Edited by Saso Dzeroski and Nada Lavrac Address all communication related to the ILP Newsletter to ilpnet@ijs.si To subscribe/unsubscribe send email with subject SUBSCRIBE/UNSUBSCRIBE ILPNEWS Send contributions in messages with subject heading ILPNEWS CONTRIBUTION Send comments and suggestions under subject heading ILPNEWS COMMENTS Back issues of the Newsletter and other information about ILPNET and ILP available via the World Wide Web (WWW), URL http://www-ai.ijs.si/ilpnet.html %------------------------------------------------------------------------------% Contents: - Workshop report: MLNET Workshop on ILP for KDD at ICML'96 - Call for papers: Special issue of the journal Applied AI on ILP for KDD - Call for papers: Workshop on Frontiers of ILP at IJCAI'97 - Call for papers: ILP'97 - The Seventh International Workshop on ILP - Preliminary announcement: Summer School on ILP and KDD - Preliminary announcement: COMPULOG Meeting - Computational Logic and ML - Two positions in Machine Learning/Data Mining at GMD %------------------------------------------------------------------------------% %------------------------------------------------------------------------------% Summary of the Workshop on ILP for KDD by Bernhard Pfahringer and Johannes Fuernkranz http://www.ai.univie.ac.at/ilp_kdd/ilp-kdd-report.html The MLNet Familiarization Workshop on "Data Mining with Inductive Logic Programming (ILP for KDD)" took place on July 2nd 1996 in Bari in conjunction with ICML96. Its was focussed on the potential of ILP as a tool for data mining and its shortcomings. Many standard methods for Knowledge Discovery in Databases (KDD) are constrained to processing a single relational table, whereas many real-world databases are structured into several tables containing interrelated information. As Inductive Logic Programming (ILP) algorithms explicitly aim at exploiting structured information, KDD should be a fruitful research and application area for ILP. Yet, ILP algorithms seem to be rarely used in KDD. The papers presented were grouped into applications, algorithmical developments, and database interface issues. The papers and other information on the workshop can be downloaded from http://www.ai.univie.ac.at/ilp_kdd/. The applications section consisted of three papers. Uros Pompe presented a paper which demonstrated a successful application of learning a two-voice counterpoint from a musical database with a stochastic ILP algorithm. Saso Dzeroski gave an interesting talk on a chemical application, where the relational learning algorithm FOIL performed significantly worse than propositional learning systems. The reason was FOIL's tendency to overfit the data even though it learned simpler rules. The most promising results on this task were achieved by a relational instance-based learning system. The paper by David Lorenzo, who unfortunately was not able to attend the workshop, describes a technique for applying ILP algorithms to a wide class of temporal databases. He demonstrated his approach with an application where he used CLAUDIEN on a clinical database. The algorithms section contained four presentations. The first talk argued that conventional ILP algorithms might be too inefficient for KDD and suggested instead the use of propositional inverse resolution for restructuring databases. In the next talk Wim Van Laer presented techniques for extending ICL to handle numeric data and multi-class learning problems. Markus Wiese suggested that FOIL-like algorithms can gain efficiency and avoid myopia by using a bi-directional search starting from random clauses. Finally Gianni Semeraro defined a refinement operator that is ideal for a subclass of Datalog clauses under theta-subsumption and demonstrated its effectiveness in an application of electronic document classification. The last section was on integrating ILP algorithms with database systems. It contained two complementary presentations, in which it was shown how to interface the ILP algorithms RDT and CLAUDIEN with real databases. The authors of both papers agreed that the restriction of most ILP systems to have all relevant background knowledge in main memory has to be overcome in order to tackle larger KDD tasks. They also found that a one-to-one mapping from relational tables in a database to PROLOG relations is impractical and suggested several alternatives. The general discussion concluding the workshop was dominated by the quest of defining data mining itself. A pragmatic definition based on the size of the input description was dismissed by the majority of the attendants, but also the somewhat ad-hoc definition of data mining as an interactive process is controversial. Interestingly, nobody tried to define data mining or Knowledge Discovery in Databases (KDD) by its presumed result, i.e., are we able to (re)discover some interesting knowledge from the given data. The least common denominator agreed to seemed to be "ILP might serve well as one of the tools needed for data mining" (the practitioner's view as expressed by Gholamreza Nakhaeizadeh). The remainder of the discussion centered on the question what sort of tools ILP can provide for KDD along the dimensions propositional vs. relational learning, optimizing predictive accuracy vs. discovery of explicit knowledge, classification vs. general discovery, extensional vs. intensional background knowledge, relational dabases vs. relational learning and efficiency vs. complexity. The general conclusion seemed to be that one of the main advantages of ILP is its flexibility in incorporating various forms of background knowledge. In particular the ability of many ILP systems to use strong language biases could proof invaluable for large KDD tasks. %------------------------------------------------------------------------------% %------------------------------------------------------------------------------% 1st Call For Papers Applied Artificial Intelligence Special issue on First-Order Knowledge Discovery in Databases (URL: http://www.ai.univie.ac.at/ilp_kdd/aai-si.html) Knowledge Discovery in Databases (KDD) is the non-trivial process of identifying valid, novel, potentially useful, and ultimately understandable patterns in data (Fayyad, Piatetsky-Shapiro & Smyth, 1996). Machine Learning algorithms form the core of many KDD systems and applications. However, standard inductive learning techniques are constrained to processing a single relational table, whereas many real-world databases are structured into several tables containing interrelated information. Learning algorithms that are able to use representations in first-order logic, in particular Inductive Logic Programming (ILP) algorithms, explicitly aim at exploiting structured information. Thus KDD is a fruitful research and application area for ILP. A recent MLnet Workshop, held at the ICML-96, focussed on a discussion of the potential contribution of ILP for KDD. Information on the workshop including a short summary and all accepted papers can be found at http://www.ai.univie.ac.at/ilp_kdd/. The general conclusion was that ILP can be a valuable tool for data mining, its main advantages being the expressiveness of first-order logic as a representation language and the ability of many ILP systems to use strong language biases for restricting the huge search space. ILP has a high flexibility in incorporating various forms of background knowledge, which can be invaluable for large KDD tasks. The special issue on "First-Order Knowledge Discovery in Databases" of the Applied Artificial Intelligence Journal will thus welcome papers that focus on one or more of the following topics: * Embedding ILP into the KDD process * Necessary pre- and post-processing steps for real-world applications * Interfacing ILP systems with database managers * Scalability of ILP for real-world databases * Criteria for quantifying the complexity of ILP problems * Evaluation of gain and price of ILP versus propositional learning * Non-classification learning and discovery in a first-order framework * Benefits of using background knowledge and/or strong explicit biases * Innovative real-world applications of ILP Papers on related subjects are also welcome, but a strong focus on applications and database issues is required for all submissions. Submissions Papers should be prepared according to usual standards for journal submissions. The approximate length of a manuscript should be between 8,000 to 10,000 words. Final manuscripts of accepted papers will have to be formatted according to the Instructions to Authors, which can be found in all issues of the journal. Authors have to submit four copies of their manuscripts to * Johannes Fuernkranz * Austrian Research Institute for Artificial Intelligence * Schottengasse 3 * A-1010 Vienna * AUSTRIA * juffi@ai.univie.ac.at or * Bernhard Pfahringer * Department of Computer Science * University of Waikato * Hamilton * NEW ZEALAND * bernhard@cs.waikato.ac.nz whichever is more convenient. Submission Deadline: April 30, 1997 Each paper submitted for publication will be judged by its originality, adequacy of method, significance of findings, and relevance to the special issue's subject matter. It should be as concise as possible, yet sufficiently detailed to permit critical review. Each manuscript must be accompanied by a statement that it has not been submitted or published elsewhere. The authors of accepted papers will be asked to transfer the copyright to the publisher. %------------------------------------------------------------------------------% %------------------------------------------------------------------------------% CALL FOR PARTICIPATION and PAPERS IJCAI-97 Workshop on FRONTIERS OF INDUCTIVE LOGIC PROGRAMMING Monday 25 August 1997 GENERAL INFORMATION The IJCAI-97 one day workshop on "Frontiers of ILP" in Nagoya, Japan, will take place on August 25, immediately prior to the start of the main IJCAI conference. TECHNICAL DESCRIPTION Inductive logic programming (ILP) is a recent subfield of artificial intelligence that studies the induction of first order formulae from examples. The purpose of this workshop is twofold: on the one hand, we wish to widen the scope of ILP by investigating its relations to neighboring fields, and on the other hand, we wish to make ILP more accessible for researchers from neighboring fields. The workshop therefore solicits papers that lie at the frontiers of ILP with neighboring fields. A non-exclusive list of interesting topics for the workshop includes : * ILP and Software Engineering: what has ILP to offer to Software Engineering ?, and in what way can Software Engineering help to design ILP systems and applications ? * ILP for Knowledge Discovery in Databases : ILP aims at learning complex rules involving multiple relations from small databases, whereas KDD typically induces simple rules about a single relation from a large database. Furthermore, ILP allows to exploit background knowledge in a variety of ways. Can KDD and ILP be succesfully combined ? * ILP and Computational or Algorithmic Learning Theory : though many results have been obtained concerning the learnability of inductive logic programming, most of the results are negative and most of the positive results are reducible to propositional learning methods. Is there a mismatch of COLT with ILP ? and if so, what can be done about it ? * ILP versus propositional learning methods : Since the very start of ILP, researchers and practioners of machine learning have wondered about the relation between ILP and propositional learning methods. Theoretical and experimental questions that arise include: when to use ILP and when to use propositional learning methods ? under what circumstances can ILP be reduced to propositional learning ? what is the price to pay for using first order logic in terms of efficiency ? * ILP and Knowledge Representation : ILP has traditionally employed computational logic to represent hypotheses and observations. Alternative well-founded knowledge representation formalisms have received little attention (with the exception of CLASSIC). What can ILP learn from Knowledge Representation ? and in what well-founded Knowledge Representation formalisms is induction feasible ? * ILP in multistrategy learning : Multistrategy learning combines multiple learning strategies. What role can ILP play for multistrategy learning ? * ILP and Probabilistic reasoning: in contrast to propositional learning methods, ILP has not used probabilistic representations. How can ILP incorporate such representations ? and how can it interact with methods such as Bayes nets or Hidden Markov Models ? * ILP for Intelligent Information Retrieval: The rapid development of the World Wide Web has spawned significant interest in intelligent information retrieval. In particular, the need for algorithms for reliably classifying textual documents into given categories (like interesting/uninteresting) be useful for a wide variety of tasks. Currently, most learning algorithms are not able to make use of structural information like word order, succesive words, structure of the text, etc. Can ILP algorithms offer advantages over conventional information retrieval or machine learning algorithms for this sort of tasks? * Applications of ILP in subfields of AI : ILP has been applied to other subfields of AI, including natural language processing, intelligent agents and planning. Further applications of ILP within AI are solicited. Both position papers about the relation of ILP to other fields, as well as research papers that make specific techical contributions are solicited. However, to stimulate discussion, it is expected that each technical paper also clarifies the position of ILP with regard to the neighboring field(s) it addresses. Except for the presentation of position and technical papers, the workshop will also feature a panel discussion on the frontiers of ILP and possibly an invited talk. ORGANISERS Luc De Raedt (chair and primary contact) Saso Dzeroski Koichi Furukawa Fumio Mizoguchi Stephen Muggleton PROGRAMME COMMITTEE Francesco Bergadano (Italy) Luc De Raedt (co-chair, Belgium) Saso Dzeroski (Slovenia) Johannes Furnkranz (Austria) Koichi Furukawa (Japan) David Page (U.K.) Fumio Mizoguchi (Japan) Ray Mooney (U.S.A.) Stephen Muggleton (co-chair, U.K.) CALL FOR PARTICIPATION Participation is open to all members of the AI Community. However, to encourage interaction and a broad exchange of ideas the number of participants will be strictly limited (preferably under 30 and certainly under 40). Participants will be selected on the basis of submissions. Three types of submissions will be considered : 1) technical contributions (ideally, a 3 to 5 page extended abstract, in the IJCAI Proceedings Format, 3000-4000 words), 2) position papers (ideally, a 1 to 3 page abstract in the IJCAI Proceedings Format, 1000 - 3000 words) 3) a statement of interest (ideally, a one page motivation of why you would like to participate, 300- 500 words) Only submissions of type 1) and 2) will be considered for presentation at the workshop and inclusion in the workshop notes. Submissions should be received no later than April 1, 1997, and must include first author's complete contact information, including address, email, phone, and fax number. Though 1 April is the hard deadline, the authors are encouraged to submit their material by 24 March, in order to facilitate the reviewing process. Double submissions with the ILP-97 Workshop (which is to take place in Prague, September 1997) are allowed. SUBMISSIONS Submit papers by email (postscript) and surface mail (2 copies) to Luc De Raedt Dept. of Computer Science Katholieke Universiteit Leuven Celestijnenlaan 200A B-3001 Heverlee Belgium Email : Luc.DeRaedt@cs.kuleuven.ac.be IMPORTANT DATES - Paper submission : 1 April - Notification to Authors : 21 April - Camera ready copy : the submissions themselve will serve as camera ready copy (submissions in the IJCAI Proceedings Style are strongly preferred, see http://www.ijcai.org/ijcai-97/ for details) PUBLICATION The accepted submissions will be included in the workshop notes to be distributed at the workshop. Post-conference publication of a selection of the workshop papers will be considered and discussed at the workshop. COSTS To cover costs, a fee of $US 50 will be charged, in addition to the normal IJCAI-97 conference registration fee. Attendees of IJCAI workshops will be required to register for the main IJCAI conference. %------------------------------------------------------------------------------% %------------------------------------------------------------------------------% ILP-97 The Seventh International Workshop on Inductive Logic Programming 17-19 September 1997, Prague, Czech Republic General Information: ILP-97 is the seventh in a series of international workshops on Inductive Logic Programing. ILP-97 will be preceeded by a two-day tutorial on ILP for KDD (15-17 September 1997) and followed by a one-day meeting (20 September 1997) of the area "Computational Logic and Machine Learning" of the European Network of Excellence on Computational Logic (COMPULOG). The Proceedings of ILP-97 will be published by Springer. Program: The scientific program will include invited talks by Usama Fayyad, Georg Gottlob, and Jean-Francois Puget, as well as presentations of accepted papers. Submissions are invited that describe theoretical, empirical and applied research in all areas of ILP. This includes, for example, results concerning logical settings for ILP, learning in the context of higher-order logics and constraint logic programming, as well as ILP systems that use probabilistic techniques and heuristics. Contributions that describe the use of ILP approaches in areas such as natural language processing, knowledge discovery in databases, intelligent agents, information retrieval, etc. are encouraged. Submissions describing applications of ILP methods to real-world problems are especially welcome. Program Chairs: Nada Lavrac and Saso Dzeroski J. Stefan Institute, Jamova 39, 1000 Ljubljana, Slovenia Email: Nada.Lavrac@ijs.si, Saso.Dzeroski@ijs.si Phone: +386 61 177 3272 (Nada) or +386 61 177 3217 (Saso) Fax: +386 61 125 1038 or +386 61 219 385 Program Committee: F. Bergadano (Italy) H. Bostrom (Sweden) I. Bratko (Slovenia) W. Cohen (USA) L. De Raedt (Belgium) P. Flach (Netherlands) S. Matwin (Canada) S. Muggleton (UK) M. Numao (Japan) D. Page (UK) C. Rouveirol (France) C. Sammut (Australia) M. Sebag (France) A. Srinivasan (UK) S. Wrobel (Germany) Local Chair: Olga Stepankova, Czech Technical University, Faculty of Electrical Engineering, Technicka 2, CZ-166 27 Prague 6, Czech Republic Email: step@lab.felk.cvut.cz, Phone: +42 2 293 107 or +42 2 2435 7233 Fax: +42 2 2435 7224 Paper Submission and Important Dates: Submit full papers of max. 5000 words. Submissions should be sent to the Program Chairs in 5 copies. A title page (including title, authors, contact author's full address, email, phone, fax, as well as abstract) must be sent by email to ilp97@ijs.si. Submission deadline: 31 March 1997 Notification of acceptance: 31 May 1997 Camera ready copy: 16 June 1997 Workshop: 17-19 September 1997 URL: http://www-ai.ijs.si/SasoDzeroski/ilp97.html %------------------------------------------------------------------------------% %------------------------------------------------------------------------------% -------------------------------------------------------------- Summer School on Inductive Logic Programming and Knowledge Discovery in Databases (ILP and KDD) -------------------------------------------------------------- Prague, Czech Republic, September 15-17, 1997 -------------------------------------------------------------- Organization: * Program: Saso Dzeroski and Nada Lavrac * Local: Olga Stepankova and Dimitar Kazakov -------------------------------------------------------------- -------------------------------------------------------------- The summer school will provide its attendants with up-to-date knowledge on ILP techniques relevant for solving practical problems of KDD, as well as with existing ILP applications in practical domains. Hands-on exercises will be included in order to give end-users a feeling of what practical problems current ILP systems are capable of solving. In addition to participants from the End-user-club of the ESPRIT ILP2 Project, the school will be open to other participants as well. It will take place immediately before ILP-97, the Seventh International Workshop on Inductive Logic Programming. For more information on the Summer School check out http://www-ai.ijs.si/SasoDzeroski/ilpkdd97.html or send email to ilp97@ijs.si -------------------------------------------------------------- -------------------------------------------------------------- Preliminary program -------------------------------------------------------------- Monday, SEP 15 10:00 - 13:00 - Saso Dzeroski and Nada Lavrac: Introduction to ILP and its applications 14:30 - 18:00 - Stefan Wrobel and Dietrich Wettschereck: ILP for KDD Hands-on exercises with KEPLER, an integrated KDD tool Tuesday, SEP 16 09:00 - 12:30 - Stephen Muggleton and Ashwin Srinivasan: Explanatory ILP and its applications Hands-on exercises with PROGOL 14:30 - 18:00 - Luc De Raedt + assistant: Descriptive ILP and its applications Hands-on exercises with CLAUDIEN Wednesday, SEP 17 09:00 - 12:30 - Usama Fayyad: The KDD methodology -------------------------------------------------------------- %------------------------------------------------------------------------------% %------------------------------------------------------------------------------% --------------- CFP: COMPULOG Net Area Meeting ---------------- Representation issues in reasoning and learning Area Meeting of CompulogNet Area "Computational Logic and Machine Learning" Prague, 20 September 1997 in conjunction with the Seventh International Workshop on Inductive Logic Programming ILP-97 Call for contributions Background Logical formalisms typically allow many degrees of freedom when it comes to representing real-world knowledge. For instance, one may represent functions in the domain of discourse either by function symbols or by predicates; one may include semantic equality in the language, or be happy with unification; one may insist that predicates and constants are ontologically meaningful, in the sense that they refer to properties, relations and objects in the domain of discourse, but one may also be more liberal by admitting expressions such as "colour(X,red)", in which the property of redness is represented by a constant rather than a monadic predicate. Such choices can be approached from different perspectives. For instance, a philosopher would probably choose the first alternative in each of the issues just mentioned; a logician interested in common-sense reasoning would look for non-truthfunctional alternatives to material implication; a computer scientist would be more concerned with efficiency of algorithms processing the formalism, and would find reasons to eschew the use of semantic equality; and a machine learner might choose a particular representation scheme simply because it suits the tabular form of her data, or because it allows to express the induced knowledge in a form comprehensible for the domain expert. Focus of the Area Meeting This CompulogNet Area Meeting concentrates on the practial aspects of representation formalisms as employed in Computational Logic and Machine Learning. In Machine Learning the main criterion to choose a particular scheme for representing data or hypotheses is pragmatic: does it lead to useful results? In Computational Logic the choice for particular reasoning schemes has mostly been guided by the efficiency of proof procedures and the applicability to real-world reasoning tasks. In both cases, a fully developed methodology for choosing a particular representation scheme is lacking. Furthermore, the relationship between the choices made by computational logicians and machine learners is unclear. The intention of the Area Meeting is to discuss these and related questions, with an emphasis on the following issues: CHOOSING THE RIGHT REPRESENTATION SCHEME FOR LEARNING * Case studies: what representation scheme worked well for what learning task, and perhaps more importantly, which ones didn't? * Specific debates: propositional vs. predicate logic, Datalog vs. Prolog, Horn clauses vs. constraints, ... * Towards a representation methodology: characteristics of domains vs. characteristics of representations COMPARING REPRESENTATIONS FOR LEARNING AND REASONING * How does efficient reasoning relate to learnability? * Are there particular representation schemes that work well in reasoning but poorly in learning, and vice versa? * Does learning need reasoning, does reasoning need learning, and if so, how does this influence a choice of an effective representation scheme BEYOND CLASSICAL LOGIC FOR LEARNING Do we need: * description logic as an intermediate between propositional and predicate logic * CLP enabling sophisticated number handling * extensions to predicate logic (temporal logic or situation calculus for dynamic domains, epistemic logic for multi-agent domains, ...) Contributions and format of the Area Meeting Participants are required to submit a short position paper (2-4 pages) addressing at least one of the meeting's central issues, identified above. Electronic submission of postscript files to Peter.Flach@kub.nl is strongly preferred. Contributed position papers will be made available on the Web before the meeting. Authors of selected contributions will be asked to give a short presentation. In addition, there will be several invited contributions. There will be ample time for discussion. Important dates - 2 June 1997 - submission of position papers - 30 June 1997 - notification and provisional program - 20 September 1997 - Area Meeting Organizers Peter Flach Nada Lavrac Tilburg University J. Stefan Institute P.O.Box 90153 Jamova 39 5000 LE Tilburg 1000 Ljubljana the Netherlands Slovenia Peter.Flach@kub.nl Nada.Lavrac@ijs.si WWW: http://machtig.kub.nl:2080/AreaMeeting97/ ------------- End CFP: COMPULOG Net Area Meeting --------------- %------------------------------------------------------------------------------% %------------------------------------------------------------------------------% Two positions in Machine Learning/Data Mining at GMD GMD's FIT.KI department (the AI research division of the Institute for Applied Computer Science) is looking to fill two scientist positions (M.S./Diplom or postdoc level) in the area of Machine Learning/Data Mining. We are looking for excellent people with a strong background in one or both of these areas, preferably combining both theoretical/scientific and application/software-engineering skills. Applications at both the postdoctoral and the M.S. level are welcome. You will be working as a research scientist in one of our current ML/DM projects, KESO or ILP2, and will be part of FIT's data mining group consisting of currently 4 people. Scientific work, writing and presentation of papers, and application and software work will both be part of your job. M.S. level applicants will be given time to complete their Ph.D.s while at GMD. Both positions are to be filled as soon as possible, for a period of initially two or three years, renewable for up to five years. Salary is according to the BAT IIa tariff, in the range of approx. DEM 50.000 to DEM 80.000 depending on age, qualifications, and marital status. For more information about FIT.KI, see http://nathan.gmd.de, for more information about the ML/data mining group, see http://nathan.gmd.de/projects/ml/home.html. If you are interested in such a position, please send your application material to Dr. Stefan Wrobel GMD, FIT.KI Schloss Birlinghoven 53754 Sankt Augustin Germany stefan.wrobel@gmd.de to be received no later than March 23, 1997 (preferably by paper mail, but E-Mail is o.k. if otherwise you cannot meet the deadline). Please include at least a brief curriculum vitae, description of your qualifications, research experience and future research interests, degree/grade information (if relevant) and if applicable, a selection of three of your best publications (full text copy). We are looking forward to your application! -------------------------------------------------------------- Dr. Stefan Wrobel GMD -- German Natl. Research Center for Information Technology FIT.KI, Schloss Birlinghoven, 53754 Sankt Augustin, Germany Tel.: +49/2241/14-0, Fax: -2889 E-Mail: stefan.wrobel@gmd.de WWW http://nathan.gmd.de/persons/stefan.wrobel.html Secr.: D. Boethgen Tel. -2731, E-Mail: dagmar.boethgen@gmd.de -------------------------------------------------------------- %------------------------------------------------------------------------------% From pazzani@super-pan.ICS.UCI.EDU Wed Mar 12 16:06:03 1997 Received: from lucy.cs.wisc.edu (lucy.cs.wisc.edu [128.105.2.11]) by sea.cs.wisc.edu (8.6.12/8.6.12) with ESMTP id QAA06931 for ; Wed, 12 Mar 1997 16:05:56 -0600 Received: from paris.ics.uci.edu (mmdf@paris.ics.uci.edu [128.195.1.50]) by lucy.cs.wisc.edu (8.7.6/8.7.3) with SMTP id QAA18993; Wed, 12 Mar 1997 16:05:45 -0600 (CST) Received: from super-pan.ics.uci.edu by paris.ics.uci.edu id aa24732; 12 Mar 97 9:20 PST To: ML-LIST:; Subject: Machine Learning List: Vol. 9, No. 4 Reply-to: ml@ics.uci.edu Date: Wed, 12 Mar 1997 08:59:26 -0800 From: Michael Pazzani Message-ID: <9703120920.aa24732@paris.ics.uci.edu> Machine Learning List: Vol. 9, No. 4 Wednesday, March 12, 1997 Contents: Machine Learning List: MLJ Table of Contents Special Issue on Machine Learning in Civil Engineering Abbadingo One: DFA Learning Competition IEEE Transaction on Neural Networks NIPS*97 Call for Papers GP-97 Late-Breaking Papers Call Workshop "ML for UM" CFP: Learning in Autonomous Robots Special Session on Artificial Immune Systems (CFP) CFP: LP + Multiagent Special issue CFP for Data Mining and Knowledge Discovery C5.0 SGI MineSet Available for Varsity Members Job Ad: IBM Data Mining Analysts Job offered in information extraction and learning, data mining ABERDEEN VACANCY Applied Data Mining Research Technician ML Posts at Bristol Does plasticity imply local learning? And other questions The Machine Learning List is moderated. Contributions should be relevant to the scientific study of machine learning. Mail contributions to ml@ics.uci.edu. Mail requests to be added or deleted to ml-request@ics.uci.edu. Back issues may be obtained from http://www.ics.uci.edu/AI/ML/Machine-Learning.html ---------------------------------------------------------------------- From: "Jeffrey C. Schlimmer" Date: Mon, 3 Mar 1997 11:45:34 -0800 Subject: Machine Learning List: MLJ Table of Contents Machine Learning Journal Table of Contents Vol. 26, No. 2/3 (February/March, 1997) Guest Editors' Introduction, Stephen Muggleton and David Page, Page 97. Clausal Discovery, Luc de Raedt and Luc Dehaspe, Page 99. First Order Regression, Aram Karalic and Ivan Bratko, Page 147. Learning Qualitative Models of Dynamic Systems, David T. Hau and Enrico W. Coiera, Page 177. Generalization of Clauses Relative to a Theory, Peter Idestam-Almquist, Page 213. PAL: A Pattern-Based First-Order Inductive System, Eduardo F. Morales, Page 227. __ Dr. Jeffrey C. Schlimmer, Asst. Prof., School of EE & CS, Washington State University, Pullman, WA 99164-2752, (509) 335-2399, (509) 335-3818 FAX http://www.eecs.wsu.edu/~schlimme/ PGP key: ftp://ftp.eecs.wsu.edu/pub/pgp/schlimmer.hqx, .txt powerPen Faculty Advisor, powerPen@eecs.wsu.edu http://www.eecs.wsu.edu/~schlimme/newton/index.shtml ftp://ftp.eecs.wsu.edu/pub/newton/ ------------------------------ From: Yoram Reich Date: Tue, 25 Feb 1997 09:57:27 +0200 (IST) Subject: Special Issue on Machine Learning in Civil Engineering The July 1997 issue of the journal: Microcomputers in Civil Engineering Journal of Computer-Aided Civil and Infrastructure Engineering will feature a special issue on Machine Learning that I edited. There are 6 papers in the issue whose titles are: 1. Biedermann, J. (1997) Representing Design Knowledge with Neural Networks. 2. Manevitz, L., Yousef, M. and Givoli, D. (1997) Finite Element Mesh Generation Using Self-Organized Neural Networks. 3. Pompe, P. P. M. and Feelders (1997) Using Machine Learning and Statistics to Predict Corporate Bankruptcy of Construction Companies: A Comparative Study. 4. Reich, Y. (1997), Machine Learning Techniques for Civil Engineering Problems. 5. Sanchez, M., Cortes, U., R.-Roda, I., Poch, M. and Lafuente, J. (1997) Learning and Adaptation in Wastewater Treatment Plants through Case-Based Reasoning. 6. Zhang, J. and Yang, J. (1997), An Application of Instance-Based Learning to Highway Accident Frequency Prediction. Abstract and an-almost-complete-version of my review paper are available through: http://or.eng.tau.ac.il:7777/topics/mice-ml.html Cheers, Yoram Department of Solid Mechanics, Materials and Structures, Faculty of Engineering, Tel Aviv University, Ramat Aviv 69978, Israel Tel: + 972 3 6407385, Fax: + 972 3 6407617 email: yoram@eng.tau.ac.il, http://or.eng.tau.ac.il:7777/ ------------------------------ From: Barak Pearlmutter Date: Thu, 6 Mar 97 18:40 MST Subject: Abbadingo One: DFA Learning Competition Abbadingo One: DFA Learning Competition Announcement & Call for Participation In order to encourage the development of better grammar induction algorithms, the Abbadingo One competition will award at least $1,024 to the designer of the system that is most successful at discovering the structure of random deterministic finite automata, as assessed by a graded series of nine benchmark problems. The competition ends on 15-Nov-1997. This competition is being sponsored by, among others, * The Computer Science Department at the University of New Mexico, which is providing computational support. * The Kluwer Academic journal "Machine Learning," which will give priority treatment to a paper describing the award winning algorithm. * The Santa Fe Institute, which will host the award ceremony. * The "Journal of Artificial Intelligence Research." For details retrieve http://abbadingo.cs.unm.edu/ Good luck, and may the best algorithm win! __ Competition Kevin J. Lang organizers: Barak A. Pearlmutter ------------------------------ From: Ron Sun Date: Sun, 23 Feb 1997 20:41:26 -0600 Subject: IEEE Transaction on Neural Networks Call For Papers special issue of IEEE Transaction on Neural Networks on ``Neural Networks and Hybrid Intelligent Models: Foundations, Theory, and Applications'' Guest Editors: C. Lee Giles, Ron Sun, Jacek M. Zurada Hybrid systems, the use of other intelligence paradigms with neural networks, are becoming more common and useful. In fact it can be argued that the success of neural networks has been from its ready incorporation of other information processing approaches, including pattern recognition, statistical inference, as well as symbolic processing. Some systems (especially those incorporating symbolic processing) have been known to some segments of the scientific community as high-level connectionist models. Other systems have been referred to as knowledge insertion and extraction. However, for the many applications, there exists little (1) theoretical foundation and (2) engineering methodology for effectively developing hybrid approaches. These two aspects are the topic of this special issue. Manuscripts are solicited in neural networks and hybrid models in the following areas - Theorectical foundations of hybrid models. Mathematical analysis, theories, critiques, case studies. - Models incorporating other paradigms such as AI symbolic processing, machine learning, fuzzy systems, genetic algorithms, and other intelligent paradigms within neural networks. Techniques, methodologies, and analyses. - Methodology of engineering design of hybrid systems. - Innovative and non-trivial applications of hybrid models (for example, in natural language processing, signal and image processing, pattern recognition, and cognitive modeling). Papers will undergo the standard review procedure of the IEEE Transactions on Neural Netwoks. The special issue will appear around November 1997. Prospective authors should submit six (6) copies of the completed manuscript, on or before February 28, 1997, to one of the following three guest editors: Dr. C. Lee Giles NEC Research Institute 4 Independence Way Princeton, NJ 08540, USA Phone: 609-951-2642 Fax 609-951-2482 Email: giles@research.nj.nec.com Web: http://www.neci.nj.nec.com/homepages/giles.html Prof. Ron Sun Department of Computer Science The University of Alabama Tuscaloosa, AL 35487 Phone: (205) 348-6363 Fax: (205) 348-0219 Email: rsun@cs.ua.edu Web: http://cs.ua.edu/faculty/sun/sun.html Prof. Jacek M. Zurada Electrical Engineering Department University of Louisville Louisville, KY 40292, USA Phone: (502) 852-6314 Fax: (502) 852-6807 Email: j.zurada@ieee.org Web: under construction ------------------------------ From: Michael Jordan Date: Sat, 8 Mar 97 11:30:25 EST Subject: NIPS*97 Call for Papers CALL FOR PAPERS -- NIPS*97 Neural Information Processing Systems -- Natural and Synthetic Monday December 1 - Saturday December 6, 1997 Denver, Colorado This is the eleventh meeting of an interdisciplinary conference which brings together cognitive scientists, computer scientists, engineers, neuroscientists, physicists, and mathematicians interested in all aspects of neural processing and computation. The conference will include invited talks and oral and poster presentations of refereed papers. The conference is single track and is highly selective. Preceding the main session, there will be one day of tutorial presentations (Dec. 1), and following will be two days of focused workshops on topical issues at a nearby ski area (Dec. 5-6). Major categories for paper submission, with example subcategories (by no means exhaustive), are as follows: Algorithms and Architectures: supervised and unsupervised learning algorithms, model selection algorithms, feedforward and recurrent network architectures, localized basis functions, online learning algorithms, active learning algorithms, algorithms for combining classifiers, belief networks, combinatorial optimization. Applications: handwriting recognition, DNA and protein sequence analysis, expert systems, fault diagnosis, financial analysis, medical diagnosis, music processing, time-series prediction. Artificial Intelligence: inductive reasoning, problem solving and planning, natural language understanding, hybrid symbolic-subsymbolic systems. Cognitive Science: perception and psychophysics, development, neuropsychology, cognitive neuroscience, language, human learning and memory, attention. Implementation: analog and digital VLSI, optical neurocomputing systems, novel neuro-devices, simulation tools, parallelism. Neuroscience: functional imaging, systems physiology, neural coding, synchrony, synaptic plasticity, neuromodulation, dendritic computation, calcium dynamics, inhibition, computational models. Reinforcement Learning and Control: exploration, dynamic programming, planning, navigation, robotic motor control, process control, Markov decision processes. Speech and Signal Processing: speech recognition, speech coding, speech synthesis, rapid adaptation, robust processing, auditory scene analysis, models of human speech perception. Theory: computational learning theory, statistical mechanics of learning, dynamics of learning algorithms, learning of dynamical systems, approximation and estimation theory, combining predictors, model selection, complexity theory. Visual Processing: image processing, image coding and classification, object recognition, stereopsis, motion detection and tracking, visual psychophysics. Review Criteria: All submitted papers will be thoroughly refereed on the basis of technical quality, significance, and clarity. Novelty of the work is also a strong consideration in paper selection, but to encourage interdisciplinary contributions, we will consider work which has been submitted or presented in part elsewhere, if it is unlikely to have been seen by the NIPS audience. Authors should not be dissuaded from submitting recent work, as there will be an opportunity after the meeting to revise accepted manuscripts before submitting final camera-ready copy. Paper Format: Submitted papers may be up to seven pages in length, including figures and references, using a font no smaller than 10 point. Submissions failing to follow these guidelines will not be considered. Authors are encouraged to use the NIPS LaTeX style files obtainable by anonymous FTP at the site given below. Papers must indicate (1) physical and e-mail addresses of all authors; (2) one of the nine major categories listed above, and, if desired, a subcategory; (3) if the work, or any substantial part thereof, has been submitted to or has appeared in other scientific conferences; (4) the authors' preference, if any, for oral or poster presentation (this preference will play no role in paper acceptance); and (5) author to whom correspondence should be addressed. Submission Instructions: Send eight copies of submitted papers to the address below; electronic or FAX submission is not acceptable. Include one additional copy of the abstract only, to be used for preparation of the abstracts booklet distributed at the meeting. SUBMISSIONS MUST BE RECEIVED BY MAY 23, 1997. From within the U.S., submissions will be accepted if mailed first class and postmarked by May 20, 1997. Mail submissions to: Michael Kearns NIPS*97 Program Chair AT&T Laboratories Research Room 2A-423 600 Mountain Avenue Murray Hill, NJ 07974-0636 USA Mail general inquiries and requests for registration material to: NIPS*97 Registration Conference Consulting Associates 451 N. Sycamore Monticello, IA 52310 fax: (319) 465-6709 (attn: Denise Prull) e-mail: nipsinfo@salk.edu Copies of the LaTeX style files for NIPS are available via anonymous ftp at ftp.cs.cmu.edu (128.2.206.173) in /afs/cs/Web/Groups/NIPS/formatting The style files and other conference information may also be retrieved via World Wide Web at http://www.cs.cmu.edu/Web/Groups/NIPS NIPS*97 Organizing Committee: General Chair, Michael Jordan, MIT; Program Chair, Michael Kearns, AT&T Labs Research; Publications Chair, Sara Solla, Northwestern University; Tutorial Chair, Satinder Singh, University of Colorado; Workshops Co-Chairs, Steven Nowlan, Lexicus, and Richard Zemel, University of Arizona; Publicity Chair, Anthony Bell, Salk Institute; Local Arrangements, Arun Jagota, University of California, Santa Cruz; Treasurer, Bartlett Mel, University of Southern California; Web Master, Doug Baker, Carnegie Mellon University; Government Liaison, John Moody, OGI; Contracts, Steve Hanson, Rutgers University, Scott Kirkpatrick, IBM, Gerry Tesauro, IBM. Conference arrangements by Conference Consulting Associates, Monticello, IA. NIPS*97 Program Committee: Sue Becker, McMaster University; Joachim Buhmann, University of Bonn; Tom Dietterich, Oregon State University; Michael Kearns, AT&T Labs Research (chair); Richard Lippmann, MIT Lincoln Lab; Larry Saul, AT&T Labs Research; Jude Shavlik, University of Wisconsin; Rich Sutton, University of Massachusetts; Tali Tishby, Hebrew University; Michael Turmon, Jet Propulsion Lab; Paul Viola, MIT; John Wawrzynek, UC Berkeley; Tony Zador, Salk Institute. DEADLINE FOR RECEIPT OF SUBMISSIONS IS MAY 23, 1997 - please post - ------------------------------ From: "John R. Koza" Date: Wed, 26 Feb 1997 20:01:53 -0800 (PST) Subject: GP-97 Late-Breaking Papers Call CALL (Version 1.0) FOR LATE-BREAKING PAPERS FOR GENETIC PROGRAMMING 1997 CONFERENCE (GP-97) _________________________________________________ DEADLINE: Wednesday, June 11, 1997 _________________________________________________ Papers describing late-breaking developments in the field of genetic programming are being solicited for inclusion in a special paper-bound book to be distributed to all attendees of the Genetic Programming 1997 Conference (GP-97) to be held on July 13 - 16 (Sunday - Wednesday), 1997 at Stanford University. This special book is distinct from the conference proceedings. The purpose of late-breaking papers is to provide conference attendees with information about research that was initiated, enhanced, improved, or completed after the original paper submission deadline in January, 1997. Late-breaking papers will be presented during a poster session to be held on the evening of Monday, July 14, 1997 during the GP-97 conference at Stanford University. Arrangements will also be made to enable interested persons to purchase this book from the Stanford Book Store after the conference. Late-breaking papers will be briefly examined for relevance and minimum standards of acceptability, but will not be peer reviewed in detail. Authors will individually retain copyright (and all other rights) to their late-breaking papers and should feel free to submit them (either before or after the above deadline) for publicaton by other conferences or journals. Late-breaking papers must be submitted in camera-ready form in accordance with the GP-97 format specifications that can be found at the GP-97 WWW site (see below). Late-breaking papers should be no more than 9 pages in length. Please send TWO camera-ready copies (printed with very high quality by laser printer) and the SIGNED "permission to publish" form (below) to GP-97 Late-Breaking Papers American Association for Artificial Intelligence 445 Burgess Drive Menlo Park, CA 94025.USA PHONE: 415-328-3123 ______________________________________________________ For additional information on GP-97 conference... - on the World Wide Web: http://www-cs-faculty.stanford.edu/~koza/gp97.html - via e-mail at gp@aaai.org ______________________________________________________ In cooperation with American Association for Artificial Intelligence (AAAI), Association for Computing Machinery (ACM), SIGART, and Society for Industrial and Applied Mathematics (SIAM) ______________________________________________________ PERMISSION TO PUBLISH FORM For Late-Breaking Papers at the GP-97 Conference Title of Paper: ____________________________ Author(s): _______________________________ The undersigned (hereinafter the "Author"), desiring that the paper identified above (hereinafter the "Paper") appear in a publication tentatively entitled "Late-Breaking Papers at the Genetic Programming 1997 Conference" and to be edited by John R. Koza, hereby grants non-exclusive permission to Genetic Programming Conferences Inc., a California non-profit corporation (hereinafter "GPCI"), to prepare and print the Paper, for sale throughout the world, in this publication. The Author retains copyright, right to transfer the copyright to other parties is in the future, right to use any and all portions of the Paper in future publications by the Author, all proprietary rights (patent rights, etc.), and all other rights. The Author assigns copyright to GPCI for publication of the Paper. The Author warrants that he/she is the author and/or proprietor of the Paper; that he/she has full power to make this agreement; that the Paper does not infringe upon any copyright, trademark, or patent; and that he/she has not granted or assigned any rights on the Paper to any person or entity that would interfere with this grant of permission. Authorized Signature: _____________________ Printed Name of Signer: ___________________ Date: _______________ Address: _______________________________ _______________________________________ _______________________________________ ------------------------------ From: Mathias Bauer Date: Thu, 27 Feb 1997 09:48:36 +0100 (MET) Subject: Workshop "ML for UM" Workshop "Machine Learning for User Modeling" to be held in conjunction with Sixth International Conference on User Modeling Chia Laguna, Sardinia, 2-5 June 1997 http://www.dfki.uni-sb.de/~bauer/um-ws Call for Participation ______________________ The information about the user that is available to an interactive computer system is usually limited, so that user model acquisition is a difficult problem. Classical acquisition methods like user interviews, application-specific heuristics, and stereotypical inferences often are not flexible enough. Machine Learning is concerned with the formation of models from observations. Hence, learning algorithms seem to be promising candidates for user model acquisition systems. There have been few uses of conventional machine learning in the area of user modeling, while in the related area of interface agents, learning has been much more popular. Recently, however, a growing interest in the application of machine learning techniques for user modeling purposes has been apparent. The goal of the workshop is to bring together researchers who are interested in employing machine learning methods for user modeling and individualized human-computer interaction. Papers tackling theoretical issues but grounded with reference to practical applications of machine learning in user modeling, are particularly encouraged. Topics of Interest __________________ The following are key questions that participants are encouraged to address: * What learning tasks can be identified in user modeling systems? * Are there classes of problems in user modeling that are partic- ularly well or poorly suited to the application of machine learning methods? * Are there machine learning algorithms or classes of algorithms that are particularly appropriate / not appropriate for user modeling systems? * Are there subareas of user modeling or classes of user modeling systems where machine learning can be especially useful? * In what respects does the induction of a user model differ from other induction tasks to which machine learning is typically applied, and what implications does this have for the applic- ation of machine learning in user modeling? * In the case of the description of a concrete application: Why did you choose this particular machine learning technique? How did it affect the success of your a pplication? What general conclusions can you draw from your experiences? Deadlines _________ * April 18 - deadline for submissions * May 2 - notification of authors about acceptance * May 16 - deadline for submission of final versions of accepted papers Submission details __________________ Submissions are invited that describe aspects of the usage of machine learning techniques for user modeling and take into account all or some of the topics listed above. The paper length is restricted to at most 6 pages in a 12pt font. The email address for electronic submission of PostScript versions (the preferred way of submission) is bauer@dfki.uni-sb.de If you have to submit a hard copy, please send 4 copies of your paper to Mathias Bauer DFKI Stuhlsatzenhausweg 3 D - 66123 Saarbruecken Germany Workshop Proceedings ____________________ All accepted papers will be made available on this web site before the workshop and as a joint research report of the institutes with which the organizers are affiliated. In addition we would like to make these papers and a summary of the workshop discussions available as online proceedings. Please try to assure that you will be able to provide an HTML version of your final paper. Alternatively the PostScript version of your paper will be included Organizers __________ * Mathias Bauer, DFKI, Germany (bauer@dfki.uni-sb.de) * Wolfgang Pohl, GMD, Germany (Wolfgang.Pohl@gmd.de) * Geoff Webb, Deakin University, Australia (webb@deakin.edu.au) Important Note on Early Registration ____________________________________ The notification date for this workshop is later than the early registration deadline of 4 April 1997. If your attendance at the conference depends on whether your workshop submission is accepted, you need not register by that deadline: If your workshop submission is accepted, you will be allowed to register for the lower early registration fee (as long as your registration is received within 1 week of the acceptance of your submission). ------------------------------ From: Henry H Hexmoor Date: Sat, 1 Mar 1997 12:16:08 -0500 (EST) Subject: CFP: Learning in Autonomous Robots Dear Colleague, Maja Mataric and I are guest editing a special issue of Prof. George Bekey's Autonomous Robots Journal. Please review our CFP which is also available online. http://www.cs.buffalo.edu/~hexmoor/autonomous-robots.html Call for papers Autonomous Robots Journal Special Issue on Learning in Autonomous Robots Guest editors: Henry Hexmoor and Maja Mataric Submission Deadline: August 15, 1997 Autonomous Robots is an international journal published by Kluwer Academic Publishers, Editor-in-Chief: George Bekey Current applications of machine learning in robotics explore learning behaviors such as obstacle avoidance, navigation, gaze control, pick and place operations, manipulating everyday objects, walking, foraging, herding, and delivering objects. It is hoped that these are first steps toward robots that will learn to perform complex operations ranging from folding clothes, cleaning up toxic waste and oil spills, picking up after the children, de-mining, look after a summer house, imitating a human teacher, or overseeing a factory or a space mission. As builders of autonomous embedded agents, researchers in robot learning deal with learning schemes in the context of physical embodiment. Strides are being made to design programs that change their initial encoding of know-how to include new concepts as well as improvements in the associations of sensing to acting. Driven by concerns about the quality and quantity of training data and real-time issues such as sparse and low-quality feedback from the environment, robot learning is undergoing a search for quantification and evaluation mechanisms, as well as for methods for scaling up the complexity of learning tasks. This special issue of Autonomous Robots will focus on novel robot learning applications and quantification of learning in autonomous robots. We are soliciting papers describing finished work preferably involving real manipulator or mobile robots. We invite submissions from all areas in AI and Machine Learning, Mobile Robotics, Machine Vision, Dexterous Manipulation, and Artificial Life that address robot learning. Submitted papers should be delivered by June 1, 1997. Authors intending to submit a manuscript should contact Henry Hexmoor as soon as possible to discuss paper ideas and suitability for this issue. Manuscripts should be typed or laser-printed in English (with American spelling preferred) and double-spaced. Both paper and electronic submission are possible, as described below. For paper submissions, send five (5) copies of submitted papers (hard-copy only) to: Dr. Henry Hexmoor Department of Computer Science State University of New York at Buffalo 226 Bell Hall Buffalo, NY 14260-2000 U.S.A. PHONE: 716-645-3197 FAX: 716-645-3464 For electronic submissions, use Postscript format, ftp the file to ftp.cs.buffalo.edu, and send an email notification to hexmoor@cs.buffalo.edu Detailed ftp instructions: compress your-paper (both Unix compress and gzip commands are ok) ftp ftp.cs.buffalo.edu (but check in case it has changed) give anonymous as your login name give your e-mail address as password set transmission to binary (just type the command BINARY) cd to users/hexmoor/ put your-paper send me an email notification hexmoor@cs.buffalo.edu to let me know you transferred the paper Relevant Dates: August 15, 1997 submission deadline November 15, 1997 review deadline December 1, 1997 acceptance/rejection notifications to the authors ------------------------------ From: "Dipankar Dasgupta (Faculty)" Date: Wed, 5 Mar 1997 15:57:09 -0600 (CST) Subject: Special Session on Artificial Immune Systems (CFP) CALL FOR PAPERS Special Session on "Artificial Immune Systems and Their Applications" at SMC'97 1997 IEEE International Conference on Systems, Man, and Cybernetics October 12-15, 1997. Orlando, Florida As part of the International Conference on Systems, Man, and Cybernetics, a special session is planned on Artificial Immune Systems and Their Applications. This interdisciplinary session will be the second event in this field, after the highly successful workshop on "Immunity-Based Systems" held in Japan on December 10, 1996. The session will be devoted to exploring different immunological mechanisms and their relation to information processing and problem solving. The scope of the special session will include the following topics (but are not limited to): * Computational methods based on Immunological principles * Artificial Immune systems for Pattern Recognition * Immunity-based system for Anomaly or fault detection * Immune Network Models and their applications * Immunity-based system as a multi-agent system * Multi-agent approach for modeling and simulating immune systems * Immunity-based systems for self-diagnosis and self-organization * Immunity-based approach for collective intelligence * Immunity-based systems for optimization and search * The immune system as a prototype of Autonomous Decentralized Systems * Immunity-based approach for Artificial Life * Immunity-based approach for security of information systems * Immunological approach against computer viruses and internet worms * The immune system as a metaphor for computer based learning systems * Immunity-based systems as a distributed learning system * Immunological Computation for data mining Prospective authors are invited to submit papers related to the listed topics. Those who are interested should send a title and an extended abstract (not more than 300 words) on or before March 10, 1997 via email or air mail to the address below: Session Chair: Dipankar Dasgupta Computer Science Division Mathematical Science Dept. The University of Memphis Memphis, TN 38152-6429. Tel: (901) 678-4147 Fax: (901) 678-2480 Email:dasgupta@mathsci.msci.memphis.edu Home Page: http://www.msci.memphis.edu/~dasgupta Important Dates: March 10, 1997 Extended Abstract Due April 20, 1997 Notification of Acceptance June 15, 1997 Final Submission For general information of SMC'97, please visit the web page of the conference at http://www.rpi.edu/~smc97. ------------------------------ From: "Dr. Phan Minh Dung" Date: Sat, 8 Mar 1997 18:55:54 +0700 Subject: CFP: LP + Multiagent International Workshop on Logic Programming and Multi-Agents sponsored by the EC activity KIT011-LPKRR and the CLN in conjunction with ICLP'97, Leuven, Belgium July 8-12, 1997 To understand the paradigm shift in computing from stand-alone computer systems to distributed systems based on the client-server architecture and the internet new conceptual models for computing are needed. In particular, improved computational models of single agents and of multi-agents are necessary. We believe that computational logic and logic programming provide the useful conceptual and practical tools for developing and studying these models. The workshop will serve as a forum for disseminating research and experience in this important and fast advancing field. Topics of interest include but are not limited to: Interacting agents Meta logic programmming applied to multi agent programming Distributed reactive systems Cognitive robotics, Applications: Integration of heterogeneous autonomous systems Software agents, agent oriented interface programming, real-time systems, Relations between agent oriented programming and object-oriented programming, Inductive logic programming for self-organizing agents The main focus of the workshop is on new and original research. But we also strongly encourage the submission of papers describing products, and prototypes in development of multi-agent systems. 'Vision' papers discussing the potential of a marriage of computational logic, especially lp and multi-agents are also welcome. We also would like to have panels to discuss issues in the potential application areas of multi-agents. WORKSHOP COORDINATOR: Phan Minh Dung (Thailand) Paolo Mancarella (Italy) ORGANIZING COMMITTEE: D. DeSchreye (Belgium) P.M. Dung (Thailand) T. Kakas (Cyprus) R. Kowalski (UK) P. Mancarella (Italy) PROGRAMME COMMITTEE : L. C. Aiello (Italy) L. DeRaedt (Belgium) D. DeSchreye (Belgium) P.M. Dung (Thailand) K. Fischer (Germany) M. Gelfond (USA) R. Kowalski (UK) T. Kakas (Cyprus) P. Mancarella (Italy) C. Palamidessi (Italy) L. M. Pereira (Portugal) D. Pearce (Germany) M. Shanahan (UK) Y. Shoham (USA) J. Siekmann (Germany) F. Toni (UK) G. Wagner (Germany) IMPORTANT DATES : Submission Deadline : May, 10, 1997 Notification : June, 10, 1997 Final Version ready : July, 1, 1997 Workshop : July, 11 - 12, 1997 PAPER SUBMISSION: A PostScript file not larger than 15 pages when printed together with a separate PostScript file of the abstract not larger than 1 page should be sent to: paolo@di.unipi.it. Updated INFORMATION on the workshop is available at http://www.cs.kuleuven.ac.be/~iclp97/Multi CONTACT ADDRESS: Phan Minh Dung (dung@cs.ait.ac.th) Paolo Mancarella (paolo@di.unipi.it) ------------------------------ From: DMKDPAR Date: Thu, 27 Feb 1997 22:54:45 -0800 (PST) Subject: Special issue CFP for Data Mining and Knowledge Discovery CALL FOR PAPERS DATA MINING AND KNOWLEDGE DISCOVERY Special Issue on Scalable High-Performance Computing for KDD Guest editors: Paul Stolorz and Ron Musick ========================================== http://www.research.microsoft.com/research/datamine/dmkdpar Traditional computational techniques and computer architectures are routinely overwhelmed by the sheer volume and complexity of information generated from data-gathering instruments, computational and experimental methodologies, and business operations. The fundamental problem of extracting knowledge and insight from massive databases and datasets is shared across a wide range of fields in business, academia and government. The new field of Data Mining and Knowledge Discovery in Databases (KDD) has arisen as an interdisciplinary response to this situation, merging ideas drawn from disciplines such as statistics, pattern recognition, machine learning, databases, visualization and high performance computing. This special issue of Data Mining and Knowledge Discovery is devoted to the challenge of applying data mining and knowledge discovery methods to large, complex datasets. Implementation of data mining ideas in high-performance computing environments is crucial for coping with large-scale data. In particular, parallel and distributed systems are needed to ensure system scalability as datasets grow inexorably in size and scope. These environments include dedicated massively parallel supercomputers, super-servers built from clusters of commodity workstations and high-speed network interfaces, and heterogeneous networks distributed over regional, national and global scales. High-performance and parallel computing holds the promise of scaling to large data sets, allowing the data mining component to search a much larger set of patterns and models than traditional computational platforms and algorithms would allow. In addition, it promises to render the KDD process much more interactive by allowing fast response times for difficult search and model fitting problems. Data Mining and Knowledge Discovery, published by Kluwer Academic publishers, is the flagship publication in the rapidly growing area of KDD. In this special issue we solicit the most dramatic new developments in high performance large-scale KDD applications, highlighting the promise of the technology and identifying the main challenges for the future. Technically innovative papers that describe new theoretical developments, or tackle the application of practical data mining approaches to real problems and datasets on parallel and distributed architectures, are solicited. Topics of interest include, but are not limited to, the intersection of KDD with the following fields: Parallel implementations of datamining & KDD methods: Classification and regression: e.g. decision trees, neural nets Pattern recognition Belief nets and other Bayesian approaches Genetic programming Association rules Statistical inference Similarity detection and measurement Clustering and density estimation Change-detection Text retrieval Content-based indexing Data visualization Trend Analysis Integration of KDD techniques with scalable I/O systems: Data warehouses & federated databases Parallel file systems High-performance network interfaces Intelligent data layout Out-of-core algorithms Parallel relational querying High performance storage systems Hierarchical and distributed storage Methods to control complexity: Random sampling Anytime algorithms applied to datamining techniques New complex data-type algorithms (eg. not based on feature vectors) Domain simplification techniques Inference error/confidence characterization Parallel, clustered and/or distributed applications: Datamining on commodity-based clusters and networks Web-oriented datamining Novel applications and case studies Knowledge discovery systems and tools SCOPE AND REVIEW CRITERIA Articles are solicited that deal with both theoretic and application- oriented approaches to handling the problems inherent in large-scale KDD. All submitted articles should be relevant to KDD, clearly indicating which aspect of large-scale KDD is being addressed. Papers should be clearly written and accessible to readers from several disciplines. A well-written, motivated introduction is especially important. Assumptions and limitations of the methods described must be discussed. Contributions must represent either a fundamental advance in algorithms and methods, or a novel application with clear roots in systematic principals. The scaling properties of algorithms and architectures with respect to problem size and complexity must be discussed, and where appropriate analysis of the throughput and latencies of the systems described. In addition to full-length papers (see below), short application summaries (1-3 pages) are also encouraged. All submissions will be reviewed on the basis of relevance, originality, significance, soundness and clarity. At least three referees will review each submission independently. Results of the review will be sent to the first author via email, unless otherwise requested. SUBMISSION INSTRUCTIONS Electronic submissions are STRONGLY ENCOURAGED. Postscript copies of papers may be emailed to dmkdpar@aig.jpl.nasa.gov. Latex style files and related instructions can be obtained at the web site http://www.research.microsoft.com/research/datamine. Submissions of full papers should be limited to at most 28 pages in 12pt font, 1.5 line-spacing. Electronic submissions will speed the review process significantly, however due to Kluwer requirements, authors must also submit hardcopy papers. All authors must submit (6) hardcopy papers as follows: five (5) hardcopies to: Ms. Karen Cullen, DATA MINING AND KNOWLEDGE DISCOVERY Editorial Office, Kluwer Academic Publishers, 101 Philip Drive, Norwell, MA 02061 phone 617-871-6600 fax 617-871-6528 email: kcullen@wkap.com one (1) hardcopy to: Dr Paul Stolorz Attn: DMKD Special Issue MS 525 3660 Jet Propulsion Laboratory 4800 Oak Grove Drive Pasadena CA 91109 USA In addition, an email message containing title, abstract, and keywords must be sent to dmkdpar@aig.jpl.nasa.gov and cc-ed to kcullen@wkap.com. Please use the electronic template available on the web. For those with no network access, please call Ms. Cullen with a request at 617-871-6600. The journal emphasizes fast dissemination of results and minimal backlogs in publication time. An electronic server will be made available by Kluwer containing accepted articles and will be accessible by subscribers to the journal. Authors are encouraged to make their data available via the journal web site, allowing papers to have an "electronic appendix" containing data and algorithms. =============== IMPORTANT DATES =============== ************************************** SUBMISSION DEADLINE: May 8, 1997 ACCEPTANCE NOTIFICATION: June 20, 1997 ************************************** Enquiries about the submission process and scope of the special issue may be sent to dmkdpar@aig.jpl.nasa.gov. ------------------------------ From: Ross Quinlan Date: Wed, 5 Mar 1997 23:28:22 -0500 (EST) Subject: C5.0 Successor to C4.5 _________________ I have developed a new inductive program called C5.0. Its main advantages are: * new, faster methods for generating rules * support for boosting * optional non-uniform misclassification costs Further information and free demonstration versions are available from http://www.rulequest.com Ross Quinlan ------------------------------ From: Ronny Kohavi Date: Wed, 26 Feb 1997 21:20:35 -0800 Subject: SGI MineSet Available for Varsity Members Silicon Graphics' MineSet Available to Varsity Members ____________________________ MineSet(TM) version 1.1 is the second release of SGI's product for data mining and exploratory data analysis. MineSet integrates tools for data access, transformations, analytical data mining, and visual data mining. See http://www.sgi.com/Products/software/MineSet for more information. In addition to 30-day free evaluation copies available to any site, with the new release of SGI's Varsity program CDs (happening now), varsity members can get PERMANENT MineSet licenses. Any educational institution is eligible. To qualify, the institution must have an infrastructure capable of handling technical software support for its Silicon Graphics users who have purchased Varsity Program software packages. THE VARSITY PROGRAM AGREEMENT MUST BE COMPLETED AND SIGNED BY THE INSTITUTION AND APPROVED BY SILICON GRAPHICS. The institution buys the right to distribute Varsity Program Developer Package right-to-use licenses in multiples of 10 or 25. These licenses are maintained by purchasing yearly support. Thus, the cost of ownership is significantly reduced in the second year and beyond. How Does this Work __________________ SGI Varsity sites will get Varsity CD-ROMs with MineSet or they can download it directly from http://www.sgi.com/Products/Evaluation/evaluation.html To get a permanent license, the site administrator can use the VPX (varsity ID) number to get a license from http://www.sgi.com/Products/license.html (click the radio button for varsity). See http://www.sgi.com/silicon_campus/varsity.html for more information about the SGI's varsity program. For questions about MineSet, send e-mail to mineset@postofc.corp.sgi.com or visit our site at: http://www.sgi.com/Products/software/MineSet ------------------------------ From: gjohn@almaden.ibm.com Date: Tue, 4 Mar 1997 01:47:42 -0800 Subject: Job Ad: IBM Data Mining Analysts IBM DATA MINING ANALYST POSITIONS (please post/redistribute) Help! We're drowning in work! IBM needs 10 more analysts for its highly successful data mining group. Join our team of high-caliber PhD's in an exciting multi-faceted career in data mining: * Analyze data for customers using IBM's industry-leading data mining products * Interact directly with senior management at Fortune 500 companies * Teach data mining classes to our customers and develop course materials * Travel, see the world! (One member of our team just got back from Paris, another is heading to Australia for two weeks... these are not vacations, it's their job!) * Interact with researchers and product developers, discuss ideas for new data mining algorithms, new visualizations, and new features for our products * Assist sales reps in customer visits, be the "technical person" to answer hard questions * Work with the marketing group to help develop brochures, etc. * Attend trade shows and conferences, learn more about the industry and talk to customers * Use SQL/AWK/PERL/SAS to process data (ooh, the excitement!) The ideal candidate * has an excellent understanding of the data analysis process and has participated in several projects * is strongly technically proficient in at least some areas of data mining (background in statistics, machine learning, neural nets, or pattern recognition, or related), with a desire to learn more * has excellent communication and presentation skills * is a self-starter, good at quickly becoming a productive member of a team * is a fast learner, can quickly become an expert in a new industry and work with IBM consultants to productively apply data mining * has some unix skills, knows enough AWK and PERL to be self-sufficient in processing data * has a good sense of humor, fun to work with, enjoys taking co-workers out to dinner, insists on paying every time, etc... Positions are available for both senior applicants (professors, PhD's, MBA's, or 4+ years relevant business experience) and more junior members (MS, BS, less job experience). Salaries are competitive, and based on experience. The jobs are focused on business, but some amount of time spent on research may be negotiated. IBM's data mining group is growing quickly, and offers excellent career opportunities. For more information on data mining at IBM, see the webpage for IBM Global Business Intelligence Solutions (our parent organization) at http://www.ibm.com/bi Send resume to George H. John, gjohn@almaden.ibm.com. ASCII (plain text) via email is *strongly* preferred. Please put "DMJOBS-97:" then your name in the subject. Hardcopy may be sent to George H. John IBM Alamden Research Center 650 Harry Rd / D2 San Jose, CA 95120-6099 FAX: 408-927-2100 (put "Attn: George John" on cover sheet) IBM is an equal opportunity employer. ------------------------------ From: Peter Norvig Date: Fri, 28 Feb 1997 15:33:40 -0800 Subject: Job offered in information extraction and learning, data mining Junglee is looking for full-time employees and summer interns to work on information discovery and data mining from text documents. We're looking for creative hard-working people with experience in some of the following: databases, information extraction, parsing, regular expressions, language design, statistics, machine learning, and GUI design. Junglee develops Internet and Intranet information technology for the future and pushes it to market today. Technology that raises eyebrows and drops barriers. Founded in 1996 by four PhD students from the Stanford University Computer Science Department and a Silicon Valley veteran, Junglee Corporation has excellent funding, high-profile customers, and a strong revenue plan. Our Virtual DataBase (VDB) engine is fueled by our ability for data source description, extraction, and attribute mapping. Imagine capturing data from hunders of disparate unstructured web sites, mixing that with data from other heterogeneous, distributed database and non-database sources and turning it all into a relational aggregate with the power of full SQL queries and the ease and portability of HTML user interfaces. We call these applications PALs - powerful information sites where people can ask for and get an answer. Several of our PALs are up on the web today at www.junglee.com and www.washingtonpost.com; we are currently building more of them for some well-known companies. One of the key aspects of the technology is discovering/mining information from text. The project is lead by Peter Norvig who has done extensive work on Natural Language Processing, Machine Learning, and other Artificial Intelligence problems. While this project involves significant ground-breaking research, it is definitely a development project, not just research. Please send responses to jobs@junglee.com or by fax to 408-522-9470 and mention this posting. __ Peter Norvig norvig@junglee.com Junglee Corporation phone: 408-522-9482 1250 Oakmead Parkway fax: 408-522-9470 Suite 310 http://www.junglee.com Sunnyvale CA 94086 http://www.norvig.com ------------------------------ From: Derek Sleeman Date: Sun, 2 Mar 1997 14:57:08 GMT Subject: ABERDEEN VACANCY Announcement of Post University of Aberdeen Chair of Computing Science Applications are invited for the post of Professor of Computing Science. The new Professor will play a key role in strengthening the teaching and research activities of the Department of Computing Science. The new Professor will provide academic leadership in the development of the Department's existing areas of interest, Artificial Intelligence and Databases. Candidates should have an international reputation with an excellent record of innovative research as measured by publications and grant income. Applications from academics, research managers and others from Industry and public sector Institutions will be considered. Further, as the University of Aberdeen has recently made a major research investment in the Institute of Medical Sciences, it would be an advantage if the person had experience of working with Medical/Healthcare professionals. The person appointed will be expected to acquire a significant role in the management of the Department. Informal enquiries may be directed to Professor A R Forrester, Vice-Principal and Dean of the Faculty of Science and Engineering: Email: a.r.forrester@admin.abdn.ac.uk Tel: +44 (0)1224 272081 Fax: +44 (0)1224 272082 More details of the Department's research activities can be found on our research pages at http://www.csd.abdn.ac.uk/research/index.html or contact Professor Derek Sleeman, Head of Department: Email: dsleeman@csd.abdn.ac.uk Tel: +44 (0)1224 272295/6 Fax: +44 (0)1224 273422 For further particulars of this post, see: http://www.csd.abdn.ac.uk/people/chair_fp.html ------------------------------ From: <72420.1134@compuserve.com> Date: 28 Feb 97 11:37:40 EST Subject: Applied Data Mining Research Technician Applied Data Mining Research Technician Large corporation located in the Midwest is seeking an individual who will conduct applied data mining research on internal sources with applications to insurance marketing, underwriting, and claims-related areas. Specific responsibilities include developing and supporting various data mining applications, including the creation and interpretation of both neural-network predictive models and decision-tree segmentation models. Additional duties may include survey construction/data collection, manipulating project data via computer programming, and the preparation and presentation of project findings. Candidates should possess a background in a research-driven academic discipline, with a strong emphasis in artificial intelligence and the application of data mining techniques and procedures. Applicants should also have experience using SAS, SPSS, or an equivalent programming/statistical tool, as well as the ability to communicate effectively, both verbally and through writing, with all levels of management. Additional experience with data warehousing principles and methods and/or relational databases is desirable. Salary will be commensurate with the applicant's background and experience. Interested applicants should e-mail their qualifications and salary history to 72420,1134@compuserve.com for consideration. ------------------------------ From: Christophe Giraud-Carrier Date: Thu, 27 Feb 1997 15:38:06 +0000 () Subject: ML Posts at Bristol ***** MACHINE LEARNING AT THE UNIVERSITY OF BRISTOL ***** We bring your attention to the following advertisement of 4 new Lecturerships (equivalent to Assistant Professorships) at the University of Bristol. We would particularly like to hear from candidates with research interests in Machine Learning and related AI fields. More information on our Machine Learning Research Group may be found at: http://www.cs.bris.ac.uk/Research/MachineLearning/ http://www.cs.bris.ac.uk/~cgc/ Also, feel free to contact Christophe Giraud-Carrier (cgc@cs.bris.ac.uk) for details on current research activities at Bristol. DEPARTMENT OF COMPUTER SCIENCE UNIVERSITY OF BRISTOL LECTURESHIPS IN COMPUTER SCIENCE The Department is continuing its rapid growth and intends to appoint four new lecturers. New research themes are being developed, linked with innovative teaching at undergraduate and postgraduate levels. The department has a lively research culture and was rated grade 5A in the 1996 Research Assessment Exercise. Its major research activities are currently in computer architecture, multimedia (including vision, image processing and graphics), declarative programming and machine learning. The department has strong links with the computer, communications, microelectronics and media industries. It is a partner in PACT, a research centre in computer architecture and software, and hosts the Safety Systems Research Centre. A major new initiative on multimedia computing and content creation, supported within the Technology Foresight Programme, will start in April 1997. Informal enquiries can be obtained from Professor David May by email (dave@cs.bris.ac.uk) or telephone +44 (0)117 954 5134. For further details and an application form telephone +44 (0)117 925 6450. minicom +44 (0)117 928 8894 or e-mail Recruitment@bris.ac.uk (stating postal address ONLY) quoting reference F224. The closing date for applications is 20th March 1997. Christophe Giraud-Carrier Department of Computer Science University of Bristol The Merchant Venturers Building Woodland Road Bristol, BS8 1UB England ------------------------------ From: Asim Roy Date: Thu, 27 Feb 1997 23:20:31 -0500 (EST) Subject: Does plasticity imply local learning? And other questions I thought it might be productive to the discussion if I compiled and posted the responses I have received so far. I am posting them without any comments. I hope this will promote further discussion on this topic. The original posting is attached below for reference. ============================================================ Response # 1: Re: your note on plasticity and learning Asim, As you mention, neuroscience tends to equate network plasticity with learning. Connectionists tend to do the same. However this raises a problem with biological systems because this conflates the processes of development and learning. Even the smartest organism starts from an egg, and develops for its entire lifespan - how do we distinguish which changes are learnt, and which are due to development. No one would argue that we *learn* to have a cortex, for instance, even though it is due to massive emryological changes in the central nervous system of the animal. This isn't a problem with artificial nets, because they do not usually have a true developmental process and so there can be no confusion between the two; but it has been a long-standing problem in the ethology literature, where learnt changes are contrasted with "innate" developmental ones. A very interesting recent contribution to this debate is Andre Ariew's "Innateness and Canalization", in Philosophy of Science 63 (Proceedings), in which he identifies non-learnt changes as being due to canalised processes. Canalization was a concept developed by the biologist Waddington in the 40's to describe how many changes seem to have fixed end-goals that are robust against changes in the environment. The relationship between development and learning was also thoroughly explored by Vygotsky (see collected works vol 1, pages 194-210). I'd like to see what other sorts of responses you get, Joe Faith Evolutionary and Adaptive Systems Group, School of Cognitive and Computing Sciences, University of Sussex, UK. ================================================================= Response # 2: I fully agree with you, that local learning is not the one and only ultimate approach - even though it results in very good learning for some domains. I am currently writing a paper on the competitive learning paradigm. I am proposing, that this competition that occurs e.g. within neurons should be called local competition. The network as a whole gives a global common goal to these local competitors and thus their competition must be regarded as cooperation from a more global point of view. There is a nice paper by Kenton Lynne that integrates the ideas of reinforcement and competition. When external evaluations are present, they can serve as teaching values, if nor the neurons compete locally. @InProceedings{Lynne88, author = {K.J.\ Lynne}, title = {Competitive Reinforcement Learning}, booktitle = {Proceedings of the 5th International Conference on Machine Learning}, year = {1988}, publisher = {Morgan Kaufmann}, pages = {188__199} } Best regards, Christoph Herrmann __________________________________________________________ Christoph Herrmann Visiting researcher Hokkaido University Meme Media Laboratory Kita 13 Nishi 8, Kita- Tel: +81 - 11 - 706 - 7253 Sapporo 060 Fax: +81 - 11 - 706 - 7808 Japan Email: chris@meme.hokudai.ac.jp http://aida.intellektik.informatik.th-darmstadt.de/~chris/ __________________________________________________________ ============================================================= Response #3: I've just read your list of questions on local vs. global learning mechanisms. I think I'm sympathatic to the implications or presuppositions of your questions but need to read them more carefully later. Meanwhile, you might find very interesting a two-part article on such a mechanism by Peter G. Burton in the 1990 volume of _Psychobiology_ 18(2).119-161 & 162-194. Steve Chandler =============================================================== Response #4: A few years back, I wrote a review article on issues of local versus global learning w.r.t. synaptic plasticity. (Unfortunately, it has been "in press" for nearly 4 years). Below is an abstract. I can email the paper to you in TeX or postscript format, or mail you a copy, if you're interested. Russell Anderson ________________________________________________ "Biased Random-Walk Learning: A Neurobiological Correlate to Trial-and-Error" (In press: Progress in Neural Networks) Russell W. Anderson Smith-Kettlewell Eye Research Institute 2232 Webster Street San Francisco, CA 94115 Office: (415) 561-1715 FAX: (415) 561-1610 anderson@skivs.ski.org Abstract: Neural network models offer a theoretical testbed for the study of learning at the cellular level. The only experimentally verified learning rule, Hebb's rule, is extremely limited in its ability to train networks to perform complex tasks. An identified cellular mechanism responsible for Hebbian-type long-term potentiation, the NMDA receptor, is highly versatile. Its function and efficacy are modulated by a wide variety of compounds and conditions and are likely to be directed by non-local phenomena. Furthermore, it has been demonstrated that NMDA receptors are not essential for some types of learning. We have shown that another neural network learning rule, the chemotaxis algorithm, is theoretically much more powerful than Hebb's rule and is consistent with experimental data. A biased random-walk in synaptic weight space is a learning rule immanent in nervous activity and may account for some types of learning __ notably the acquisition of skilled movement. __________________________________________ E-mail: send request to rwa@milo.berkeley.edu Slow mail: Russell Anderson 2415 College Ave. #33 Berkeley, CA 94704 ========================================================== Response #5: Asim Roy typed ... > > B) "Pure" local learning does not explain a number of other > activities that are part of the process of learning!! .... > > So, in the whole, there are a "number of activities" that need to > be > performed before any kind of "local learning" can take place. These > aforementioned learning activities "cannot" be performed by a > collection of "local learning" cells! There is more to the process > of learning than simple local learning by individual cells. Many > learning "decisions/tasks" must precede actual training by "local > learners." A group of independent "local learners" simply cannot > start learning and be able to reproduce the learning > characteristics and processes of an "autonomous system" like the > brain. I cannot see how you can prove the above statement (particularly the last sentence). Do you have any proof. By analogy, consider many insect colonies (bees, ants etc). No-one could claim that one of the insects has a global view of what should happen in the colony. Each insect has its own purpose and goes about that purpose without knowing the global purpose of the colony. Yet an ants nest does get built, and the colony does survive. Similarly, it is difficult to claim that evolution has a master plan, order just seems to develop out of chaos. I am not claiming that one type of learning (local or global) is better than another, but I would like to see some evidence for your somewhat outrageous claims. > Note that the global learning mechanism may actually be implemented > with a collection of local learners!! You seem to contradict yourself here. You first say that local learning cannot cope with many problems of learning, yet global learning can. You then say that global learning can be implemented using local learners. This is like saying that you can implement things in C, that cannot be implemented in assembly!! It may be more convenient to implement it in C (or using global learning), but that doesn't make it impossible for assembly. Cheers, Brendan. ___________________________________________________________________ Brendan McCane, PhD. Email: mccane@cs.otago.ac.nz Comp.Sci. Dept., Otago University, Phone: +64 3 479 8588. Box 56, Dunedin, New Zealand. There's only one catch - Catch 22. =============================================================== Response #6: In regards to arguments against global learning: I think no one seriously questions this possibility, but think that global learning theories are currently non-verifiable/ non-falsifyable. Part of the point of my paper was that there ARE ways to investigate non-local learning, but it requires changes in current experimental protocols. Anyway, good luck. I look forward to seeing your compilation. Russell __________________________________________ E-mail: send request to rwa@milo.berkeley.edu Slow mail: Russell Anderson 2415 College Ave. #33 Berkeley, CA 94704 ============================================================== Response #7: I am sorry that it has taken so long for me to reply to your inquiry about plasticity and local/global learning. As I mentioned in my first note to you, I am sympathetic to the view that learning involves some sort of overarching, global mechanism even though the actual information storage may consist of distributed patterns of local information. Because I am sympathetic to such a view, it makes it very difficult for me to try to imagine and anticipate the problems for such views. That's why I am glad to see that you are explicitly trying to find people to point out possible problems; we need the reality check. The Peter Burton articles that I have sent you describes exactly the kind of mechanism implied by your first question: Does plasticity imply local learning? Burton describes a neurological mechanism by which local learning could emerge from a global signal. Essentially he posits that whenever the new perceptual input being attended to at any given moment differs sufficiently from the record of previously recorded experiences to which that new input is being compared, the difference triggers a global "proceed-to-store" signal. This signal creates a neural "snapshot" (my term, not Burton's) of the cortical activations at that moment, a global episodic memory (subject to stimulus sampling effects, etc.). Burton goes on to describe how discrete episodic memories could become associated with one another so as to give rise to schematic representations of percepts (personally I don't think that positing this abstraction step is necessary, but Burton does it). As neuroscientists sometimes note, while it is widely assumed that LTP/LTD are local learning mechanisms, the direct evidence for such a hypothesis is pretty slim at best. Of course of of the most serious problems with that view is that the changes don't last very long and thus are not really good candidates for long term (i.e., life long) memory. Now, to my mind, one of the most important possibilities overlooked in LTP studies (inherently so in all in vitro preparations and so far as I know__which is not very far because this is not my field__in the in vivo preparations that I have read about) is that LTP/D is either an artifact of the experiment or some sort of short term change which requires a global signal to become consolidated into a long term record. Burton describes one such possible mechanism. Another motivation for some sort of global mechanism comes from the so-called 'binding problem' addressed especially by the Damasio's, but others too. Somehow somewhere all the distributed pieces of information about what an orange is, for example, have to be tied together. A number of studies of different sorts have demonstarted repeatedly that such information is distributed throughout cortical areas. Burton distinguishes between "perceptual learning" requiring no external teacher (either locally or globally) and "conceptual learning", which may require the assistance of a 'teacher'. In his model though, both types of learning are activated by global "proceed-to-learn" signals triggered in turn by the global summation of local disparities between remembered episodes and current input. I'll just mention in closing that I am particularly interested in the empirical adequacy of neuropsychological accounts such as Burton's because I am very interested in "instance-based" or "exemplar-based" models of learning. In particular, Royal Skousen's _Analogical Modeling of Language_ (Kluwer, 1989) describes an explicit, mathematical model for predicting new behavior on analogy to instances stored in long term memory. Burton's model suggests a possible neurological basis for such behavior. ============================================================== Response #8: ******************************************************************* Fred Wolf E-Mail: fred@chaos.uni-frankfurt.de Institut fuer Theor. Physik Robert-Mayer-Str. 8 Tel: 069/798-23674 D-60 054 Frankfurt/Main 11 Fax: (49) 69/798-28354 Germany ******************************************************************* Dear Asim Roy, could you please point me to a few neuroBIOLOGICAL references that justify your claim that > > A predominant belief in neuroscience is that synaptic plasticity > and LTP/LTD imply local learning (in your sens). > I think many people appreciate that real learning implies the concerted interplay of a lot of different brain systems and should not even be attempted to be explained by "isolated local learners". See e.g. the series of review-papers on memory in a recent volume of PNAS 93 (1996) (http://www.pnas.org/). Good luck with your general theory of global/local learning. best wishes Fred Wolf ============================================================== Response #9: Comp-Neuro Mailing List wrote: > > =================================================================== I am into neurocomputing for several years. I read your arguments with interest. They certainly deserve further attention. Perhaps some combination of global-local learning agents would be the right choice. - Vassilis G. Kaburlasos Aristotle University of Thessaloniki, Greece ------------------------------ End of ML-LIST (Digest format) **************************************** From kacprzyk@ibspan.waw.pl Thu Mar 13 03:46:24 1997 Received: from lucy.cs.wisc.edu (lucy.cs.wisc.edu [128.105.2.11]) by sea.cs.wisc.edu (8.6.12/8.6.12) with ESMTP id DAA08150 for ; Thu, 13 Mar 1997 03:46:12 -0600 Received: from cs.uwa.oz.au (bilby.cs.uwa.oz.au [130.95.1.11]) by lucy.cs.wisc.edu (8.7.6/8.7.3) with SMTP id DAA29641 for ; Thu, 13 Mar 1997 03:46:08 -0600 (CST) Received: from (parma.cs.uwa.oz.au [130.95.1.7]) by cs.uwa.oz.au (8.6.8/8.5) with SMTP id PAA17377; Thu, 13 Mar 1997 15:48:03 +0800 Message-Id: <199703130748.PAA17377@cs.uwa.oz.au> From: Janusz Kacprzyk To: fuzzy_food@mech.ubc.ca Cc: reinforce@cs.uwa.edu.au Subject: Re: new book Date: Wed, 12 Mar 1997 12:13:37 +0100 MULTISTAGE FUZZY CONTROL A Model-Based Approach to Fuzzy Control and Decision Making Janusz Kacprzyk, Systems Research Institute, Polish Academy of Sciences, Warsaw, Poland John Wiley & Sons Chichester, New York, ... 1997 Fuzzy techniques are used to cope with imprecision in the control process. This authoritative book explains the essential principles of fuzzy logic and describes both the theoretical and practical advantages of the new model-based, prescriptive approach. Professor Kacprzyk offers a comprehensive and in depth examination of the issues underlying multistage control and decision analysis, addressing in particular fuzzy dynamic systems, fuzzy events, fuzzy probabilities and fuzzy quantifiers. The text also comprises an introduction to the basic concepts of fuzzy sets, fuzzy logic and fuzzy systems, complemented by real-world examples of the use of the model-based prescriptive approach to improve the efficiency of fuzzy control systems. Highly experienced in fuzzy control research, the author identifies new trends in the development of fuzzy sets and their direct application to decision-making processes. Fuzzy control engineers, researchers and postgraduate students will find this an ideal reference, offering a wealth of ideas for enhancing the performance of fuzzy control systems and equipping them with the tools to resolve genuine problems. Multistage Fuzzy Control is an essential handbook for those wishing to resolve real-world problems in control and decision analysis through the use of fuzzy-logic-based methods. CONTENTS Introduction; Basic Elements of Fuzzy Sets and Fuzzy Systems; A General Setting for Multistage Control under Fuzziness; Control Processes with a Fixed and Specified Termination Time; Control Processes with an Implicitly Specified Termination Time; Control Processes with a Fuzzy Termination Time; Control Processes with an Infinite Termination Time; Examples of Applications; Concluding Remarks; Index 0 471 96347 X January 1997 (cl) 338pp GBP 45.00 ORDER FORM 0471 96347 X KACPRZYK FUZZY GBP 45.00 Order TWO OR MORE books and postage is FREE! Orders for one book please add GBP 2.00. To receive order by airmail please add GBP 5.50. Phone your credit card order UK Freefone: 0800 243407 Overseas: +44 1243 770424 Fax your order form to: +44 (0) 1243 775878 Post your order form to: Dominic Rooke-Allden John Wiley & Sons Ltd, Baffins Lane Chichester, West Sussex, PO19 1UD, UK email: lkempste@wiley.co.uk Please include your postal delivery address. HOW TO PAY PLEASE TICK ONE BOX ONLY o Cheque/Payment GBP enclosed PAYABLE TO JOHN WILEY & SONS LTD o Please charge my credit/charge card o Mastercard o Visa o American Express o Diners Club o JCB Card Number / / / / / / / / / / / / / / / / / Expiry Date / / / / / Signature of cardholder Date ORDER NOT VALID UNLESS SIGNED STAY INFORMED Please keep me informed of publications in my key subject area(s) which are: ____________________________ Available from your bookseller. If you are not completely satisfied we will refund your payment without question, if books are returned in a resaleable condition within 30 days of receipt. Your order will be processed without delay, please allow 21 days for delivery. Orders for forthcoming books will be recorded and you will receive confirmation of your order. You will then be invoiced on publication. All information is subject to change without notice. Please tick here if you do not wish to receive mail from other companies SEND MY ORDER TO (PLEASE USE CAPITALS) Title & Name Job Title Department Company/University Address Post/Zip Code Country Tel Fax email Please give credit card address if different from delivery address. Purchase order enclosed Please send me/my company an invoice for prepayment EU Member states please include your VAT number: ***************************************************************** *Prof. Janusz Kacprzyk Systems Research Institute * * Polish Academy of Sciences * * Email: kacprzyk@ibspan.waw.pl ul. Newelska 6 * * Phone: + (48) (22) 36 41 03 01-447 Warsaw * * Fax: + (48) (22) 37 27 72 Poland * *****************************************************************