neuron-request@HPLABS.HP.COM (Neuron-Digest Moderator Peter Marvit) (11/16/88)
Neuron Digest Tuesday, 15 Nov 1988 Volume 4 : Issue 23 Today's Topics: Mark Jones to speak on neural nets and symbolic AI colloquia CVPR 89 submission deadline Washington Neural Network Society Technical report announcement FINAL CALL FOR PAPERS Report Available - Connectionist State Machines BBS Call For Commentators: The Tag Assignment Problem Send submissions, questions, address maintenance and requests for old issues to "neuron-request@hplabs.hp.com" or "{any backbone,uunet}!hplabs!neuron-request" ------------------------------------------------------------ Subject: Mark Jones to speak on neural nets and symbolic AI From: pratt@zztop.rutgers.edu (Lorien Y. Pratt) Organization: Rutgers Univ., New Brunswick, N.J. Date: 04 Nov 88 19:36:12 +0000 Fall, 1988 Neural Networks Colloquium Series at Rutgers Neural Nets and Symbolic AI --------------------------- Mark Jones AT&T Bell Laboratories Room 705 Hill center, Busch Campus Friday November 11, 1988 at 11:10 am Refreshments served before the talk Abstract I will present an overview of the rapidly developing area of neural networks or connectionist networks and their application to problems in Artificial Intelligence (AI). I will briefly discuss areas of perception, feature discovery, associative memory and pattern completion, but will concentrate on the relationship of neural networks to classical symbolic AI domains such as natural language processing and knowledge representation. Lorien Y. Pratt Computer Science Department pratt@paul.rutgers.edu Rutgers University Busch Campus (201) 932-4634 Piscataway, NJ 08854 ------------------------------ Subject: colloquia From: loui@wucs1.wustl.edu (Ron Loui) Organization: Washington University, St. Louis, MO Date: 04 Nov 88 21:26:44 +0000 COMPUTER SCIENCE COLLOQUIUM Washington University St. Louis 4 November 1988 TITLE: Why AI needs Connectionism? A Representation and Reasoning Perspective Lokendra Shastri Computer and Information Science Department University of Pennsylvania Any generalized notion of inference is intractable, yet we are capable of drawing a variety of inferences with remarkable efficiency - often in a few hundered milliseconds. These inferences are by no means trivial and support a broad range of cognitive activity such as classifying and recognizing objects, understanding spoken and written language, and performing commonsense reasoning. Any serious attempt at understanding intelligence must provide a detailed computational account of how such inferences may be drawn with requisite efficiency. In this talk we describe some work within the connectionist framework that attempts to offer such an account. We focus on two connectionist knowledge representation and reasoning systems: 1) A connectionist semantic memory that computes optimal solutions to an interesting class of inheritance and recognition problems extremely fast - in time proportional to the depth of the conceptual hierarchy. In addition to being efficient, the connectionist realization is based on an evidential formulation and provides a principled treatment of exceptions, conflicting multiple inheritance, as well as the best-match or partial-match computation. 2) A connectionist system that represents knowledge in terms of multi-place relations (n-ary predicates), and draws a limited class of inferences based on this knowledge with extreme efficiency. The time taken by the system to draw conclusions is proportional to the length of the proof, and hence, optimal. The system incorporates a solution to the "variable binding" problem and uses the temporal dimension to establish and maintain bindings. We conclude that working within the connectionist framework is well motivated as it helps in identifying interesting classes of limited inference that can be performed with extreme efficiently, and aids in discovering constraints that must be placed on the conceptual structure in order to achieve extreme efficiency. host: Ronald Loui ___________________________________________________________________________ 1988-89 AI Colloquium Series (through February) Sep 16 Michael Wellman, MIT/Air Force "The Trade-off Formulation Task in Planning under Uncertainty" 30 Kathryn Laskey, Decision Science Consortium "Assumptions, Beliefs, and Probabilities" Nov 4 Lokendra Shastri, University of Pennsylvania "Why AI Needs Connectionism? A Representation and Reasoning Perspective" 11 Peter Jackson, McDonnell Douglas Research Laboratories "Diagnosis, Defaults, and Abduction" 18 Eric Horvitz, Stanford University (decision-theoretic control of problem-solving) Dec 2 Mark Drummond, NASA Ames (planning) Feb 3 Fahiem Bacchus, University of Waterloo (uncertain reasoning) 10 Dana Nau, University of Maryland (TBA) other speakers to be announced ____________________________________________________________________________ ------------------------------ Subject: CVPR 89 submission deadline From: wnm@uvacs.cs.Virginia.EDU (Worthy N. Martin) Organization: U.Va. CS Department, Charlottesville, VA Date: 07 Nov 88 01:50:02 +0000 The following call for papers has appeared before, however, it is being reissued to remind interested parties of the first deadline, namely: - ----> - ----> November 16, 1988 -- Papers submitted - ----> This deadline will be held to firmly with the submission date determined by postmark. Thank you for your interest in CVPR89 Worthy Martin - ---------------------------------------------------------- CALL FOR PAPERS IEEE Computer Society Conference on COMPUTER VISION AND PATTERN RECOGNITION Sheraton Grand Hotel San Diego, California June 4-8, 1989. General Chair Professor Rama Chellappa Department of EE-Systems University of Southern California Los Angeles, California 90089-0272 Program Co-Chairs Professor Worthy Martin Professor John Kender Dept. of Computer Science Dept. of Computer Science Thornton Hall Columbia University University of Virginia New York, New York 10027 Charlottesville, Virginia 22901 Program Committee Charles Brown John Jarvis Gerard Medioni Larry Davis Avi Kak Theo Pavlidis Arthur Hansen Rangaswamy Kashyap Alex Pentland Robert Haralick Joseph Kearney Roger Tsai Ellen Hildreth Daryl Lawton John Tsotsos Anil Jain Martin Levine John Webb Ramesh Jain David Lowe Submission of Papers Four copies of complete drafts, not exceeding 25 double spaced typed pages should be sent to Worthy Martin at the address given above by November 16, 1988 (THIS IS A HARD DEADLINE). All reviewers and authors will be anonymous for the review process. The cover page will be removed for the review process. The cover page must contain the title, authors' names, primary author's address and telephone number, and index terms containing at least one of the below topics. The second page of the draft should contain the title and an abstract of about 250 words. Authors will be notified of notified of acceptance by February 1, 1989 and final camera-ready papers, typed on special forms, will be required by March 8, 1989. Submission of Video Tapes As a new feature there will be one or two sessions where the authors can present their work using video tapes only. For information regarding the submission of video tapes for review purposes, please contact John Kender at the address above. Conference Topics Include: -- Image Processing -- Pattern Recognition -- 3-D Representation and Recognition -- Motion -- Stereo -- Visual Navigation -- Shape from _____ (Shading, Contour, ...) -- Vision Systems and Architectures -- Applications of Computer Vision -- AI in Computer Vision -- Robust Statistical Methods in Computer Vision Dates November 16, 1988 -- Papers submitted February 1, 1989 -- Authors informed March 8, 1989 -- Camera-ready manuscripts to IEEE June 4-8, 1989 -- Conference ------------------------------ Subject: Washington Neural Network Society From: will@ida.org (Craig Will) Date: Mon, 07 Nov 88 15:32:23 -0500 The Washington Neural Network Society and The IEEE Computer Society Artificial Intelligence Subchapter Joint Meeting November 14, 1988 7:00 PM Speaker: C. Lee Giles Air Force Office of Scientific Research Washington, D.C. High Order Neural Networks Conventional neural networks make use of interconnections between neurons based on a single connection strength, or weight, associated with each input to a neuron. Such networks often require multiple layers of neurons to solve specific problems, and difficulties of learning and stability have not been completely solved. An alternative approach is to increase the computational power of each neuron in the network. One way to do this is to use more complex interconnections that not only include a sin- gle connection strength for each single input to a neuron, but also a weight representing the connection strength for each pair of other neurons feeding into a unit, a weight for each triple of other neurons feeding into an input, and so forth. Such networks are called higher order networks, and appear to have several advantages. They can efficiently construct high-order internal representations that can capture high- order internal representations in complex, high-dimensional data. This can allow a network to encode invariances, such as the ability to recognize a visual shape regardless of size, rotation, or a vertical or horizontal shift. Higher order networks can perform the same computations as conventional networks with fewer layers, and usually have learning times faster by orders of magnitude. In this talk Lee Giles will present an introduction to higher order neural networks and discuss their advantages over conventional approaches. He will also discuss difficulties with higher order networks, such as scaling up to large sys- tems, as well as potential solutions. Dr. C. Lee Giles is with the Air Force Office of Scien- tific Research, Bolling Air Force Base, Washington, D.C., where he is a program officer responsible for basic research on the computational aspects of neural networks. The meeting will be held at MITRE in McLean, Virginia. Take the Beltway (495) toward McLean, and get off at exit 11 (if coming from Virginia) or exit 11A (if coming from Mary- land). Take Route 123 north to the second traffic light (at Colshire Drive). Go right onto Colshire Drive and follow it to MITRE's Hayes Building at the top of the hill. From Wash- ington, take 66 West to the Beltway and go north, get off at exit 11 as above. For more information call Diane Entner at (703) 243-6996. Schedule: 7:00 - 7:10 Introduction (Craig Will, Diane Entner) 7:10 - 7:20 Neural nets at MITRE (Alexis Wieland) 7:20 - 8:20 Speaker (Lee Giles) 8:20 - 9:20 Informal discussion with refreshments ------------------------------ Subject: Technical report announcement From: David.Servan-Schreiber@A.GP.CS.CMU.EDU Date: Wed, 09 Nov 88 00:05:00 -0500 The following technical report is available upon request: ENCODING SEQUENTIAL STRUCTURE IN SIMPLE RECURRENT NETWORKS David Servan-Schreiber, Axel Cleeremans & James L. McClelland CMU-CS-88-183 We explore a network architecture introduced by Elman (1988) for predicting successive elements of a sequence. The network uses the pattern of activation over a set of hidden units from time- step t-1, together with element t, to predict element t+1. When the network is trained with strings from a particular finite- state grammar, it can learn to be a perfect finite-state recognizer for the grammar. When the net has a minimal number of hidden units, patterns on the hidden units come to correspond to the nodes of the grammar; however, this correspondence is not necessary for the network to act as a perfect finite-state recognizer. We explore the conditions under which the network can carry information about distant sequential contingencies across intervening elements to distant elements. Such information is maintained with relative ease if it is relevant at each intermediate step; it tends to be lost when intervening elements do not depend on it. At first glance this may suggest that such networks are not relevant to natural language, in which dependencies may span indefinite distances. However, embeddings in natural language are not completely independent of earlier information. The final simulation shows that long distance sequential contingencies can be encoded by the network even if only subtle statistical properties of embedded strings depend on the early information. Send surface mail to : Department of Computer Science Carnegie Mellon University Pittsburgh, PA. 15213-3890 U.S.A or electronic mail to Ms. Terina Jett: Jett@CS.CMU.EDU (ARPA net) Ask for technical report CMU-CS-88-183. ------------------------------ Subject: FINAL CALL FOR PAPERS From: Julian Dow@UK.AC.GLASGOW.VME Date: 3 Nov 88 13:47:55 To: CONNECT-BB@UK.AC.ED.EUSIP Msg ID: < 3 Nov 88 13:47:55 GMT A10120@UK.AC.GLA.VME> CALL FOR PAPERS: "BIOELECTRONICS AND BIOSENSORS" SEB MEETING: EDINBURGH, APRIL 3-7th, 1989 The programme for the meeting is being drawn up at present; a provisional list of speakers and titles is shown below: Pickard (Cardiff) "Bee brains and biosensors". Turner (Cranfield) "Biosensor design". Clark (Glasgow) "Cell patterning by contact guidance on synthetic substrates". Gross (Texas) "Simultaneous recording from multielectrode arrays". Edell (MIT) "Long term recording from neuronal implants". Pine (Caltech) "The silicon/neuron connection". Birch (Unilever) "Electrochemical sensors based on the capillary fill device". Kulys (USSR) "Biosensors based on organic metal electrodes". Pethig (Bangor) "Sensors and switches based on biological materials". Stanley (Scientific Generics) "Pitfalls and potential of the next generation of commercial biosensors". Cullen (Cambridge) "Immunosensors based on the excitation of surface plasmon polaritons on diffraction gratings". Thompson (Toronto) "The use of animal cell receptors in biosensors". A session on "computers in neurobiology" and a "CED Users' Group meeting" will run concurrently in Edinburgh. If you would like to submit a paper or poster, or are not already included on the mailing list, please reply to me below. Please return the completed form to me at the above address by the deadline of 15th November, 1988. Successful applicants will be notified shortly afterwards, and the final programme for the whole meeting, together with registration forms and accommodation booking details, will be circulated in February. ***Topics covered Developmental biology: contact guidance, studies of cell behaviour on patterned surfaces of controlled topography. In vivo Neurobiology: measurement of electrical activity in intact nervous tissue with microengineered electrode arrays. In vitro Neurobiology: "real" neural networks of cultured neurons on arrays of electrodes. Relevance to computer design. Biomedical applications: use of electrode arrays as implantable sensors and in prosthetics. Biosensors: Fundamental problems. Application to biological problems in research. Technological problems: Choice of substrate, electrode materials. Long-term stability in aqueous environments. Multiplexing and signal treatment. Data reduction. ***Interested? Please complete and return the attached form if you would like to be included on the mailing list, or write to Mail: Dr. Julian A.T. Dow, Department of Cell Biology, University of Glasgow, Glasgow G12 8QQ, Scotland. Phone: (041) 330 4616 Telex: 777070 UNIGLA Fax: (041) 330 4808 (Electronic mail address (JANET) : Julian_Dow@uk.ac.glasgow.vme) - ------------------------------------------------ SEB Conference: April 1989 To: Dr. Julian A.T. Dow, Department of Cell Biology, University of Glasgow, Glasgow G12 8QQ, Scotland. Telephone (041) 330 4616 Name: .............................................. Address: .............................................. .............................................. .............................................. .............................................. Please include me on your mailing list for further announcements on the Bioelectronics & Biosensors session of the 1989 SEB Spring Meeting in Edinburgh, on the 6th and 7th of April. I am interested in giving an oral presentation presenting a poster attending. The title of my poster presentation would be: .......................................................... My main area of interest is Cell biology Developmental Biology Neurobiology Biosensors Electronics Computing (Please circle) I am / am not a member of the SEB (you do not *need* to be a member, incidentally) Please duplicate this form and pass on to any interested colleagues. - --- End of forwarded message - --- End of forwarded message ------------------------------ Subject: Report Available - Connectionist State Machines From: rba@flash.bellcore.com (Robert B Allen) Date: Thu, 10 Nov 88 16:52:24 -0500 Connectionist State Machines Robert B. Allen Bellcore, November 1988 Performance of sequential adaptive networks on a number of tasks was explored. For example the ability to respond to continuous sequences was demonstrated first with a network which was trained to flag a given subsequence and, in a second study, to generate responses to transitions conditional upon previous transitions. Another set of studies demonstrated that the networks are able to recognize legal strings drawn from simple context-free grammars and regular expressions. Finally, sequential networks were also shown to be able to be trained to generate long strings. In some cases, adaptive schedules were introduced to gradually extend the network's processing of strings. Contact: rba@bllcore.com Robert B. Allen 2A-367 Bellcore Morristown, NJ 07960-1910 ------------------------------ Subject: BBS Call For Commentators: The Tag Assignment Problem From: harnad@confidence.Princeton.EDU (Stevan Harnad) Date: Fri, 11 Nov 88 02:32:57 -0500 Below is the abstract of a forthcoming target article to appear in Behavioral and Brain Sciences (BBS), an international, interdisciplinary journal providing Open Peer Commentary on important and controversial current research in the biobehavioral and cognitive sciences. To be considered as a commentator or to suggest other appropriate commentators, please send email to: harnad@confidence.princeton.edu or write to: BBS, 20 Nassau Street, #240, Princeton NJ 08542 [tel: 609-921-7771] ____________________________________________________________________ A SOLUTION TO THE TAG-ASSIGNMENT PROBLEM FOR NEURAL NETWORKS Gary W. Strong Bruce A. Whitehead College of Information Studies Computer Science Program Drexel University University of Tennessee Space Institute Philadelphia, PA 19104 USA Tullahoma, TN 37388 USA ABSTRACT: Purely parallel neural networks can model object recognition in brief displays -- the same conditions under which illusory conjunctions (the incorrect combination of features into perceived objects in a stimulus array) have been demonstrated empirically (Treisman & Gelade 1980; Treisman 1986). Correcting errors of illusory conjunction is the "tag-assignment" problem for a purely parallel processor: the problem of assigning a spatial tag to nonspatial features, feature combinations and objects. This problem must be solved to model human object recognition over a longer time scale. A neurally plausible model has been constructed which simulates both the parallel processes that may give rise to illusory conjunctions and the serial processes that may solve the tag-assignment problem in normal perception. One component of the model extracts pooled features and another provides attentional tags that can correct illusory conjunctions. Our approach addresses two questions: (i) How can objects be identified from simultaneously attended features in a parallel, distributed representation? (ii) How can the spatial selection requirements of such an attentional process be met by a separation of pathways between spatial and nonspatial processing? Analysis of these questions yields a neurally plausible simulation model of tag assignment, based on synchronization of neural activity for features within a spatial focus of attention. KEYWORDS: affordance; attention; connectionist network; eye movements; illusory conjunction; neural network; object recognition; retinotopic representations; saccades; spatial localization ------------------------------ End of Neurons Digest *********************