neuron-request@HPLABS.HP.COM ("Neuron-Digest Moderator Peter Marvit") (03/14/89)
Neuron Digest Monday, 13 Mar 1989 Volume 5 : Issue 13 Today's Topics: Call for Minitrack in Business Applications of NN Connection between Hidden Markov Models and Connectionist Networks EURASIP Workshop on Neural Nets NIPS POST-MEETING WORKSHOPS Preprint - Performance of a Stochastic Learning Microchip Rules and Variables in a Connectionist Reasoning System Talk at ICSI - DIFICIL Talk at ICSI - "Perceptual Organization for Computer Vision" Talk at ICSI - Spreading Activation Meets Back Propagation: Talk at ICSI - The Sphinx Speech Recognition System TR - ANNs and Sequential Paraphrasing of Script-Based Stories TR - ANNs in Robot Motion Planning TR - Dynamic Node Creating in Back-Prop Nets TR - Learning State Space Trajectories in Recurrent ANNs TR - Speeding up ANNs in the "Real World" Send submissions, questions, address maintenance and requests for old issues to "neuron-request@hplabs.hp.com" or "{any backbone,uunet}!hplabs!neuron-request" ARPANET users can get old issues via ftp from hplpm.hpl.hp.com (15.255.16.205). ------------------------------------------------------------ Subject: Call for Minitrack in Business Applications of NN From: T034360%UHCCMVS.BITNET@CUNYVM.CUNY.EDU Date: Sat, 04 Mar 89 01:22:00 -1000 CALL FOR PAPERS HAWAII INTERNATIONAL CONFERENCE ON SYSTEM SCIENCES - 23 NEURAL NET APPLICATIONS IN BUSINESS KAILUA-KONA, HAWAII - JANUARY 2-5, 1990 The Emerging Technologies and Applications Track of HICSS-23 will contain a special set of sessions focusing on a broad selection of topics in the area of Neural Net Applications in Business. The presentations will provide a forum to discuss new advances in these applications. Papers are invited that may be theoretical, conceptual, tutorial, or descriptive in nature. Of special interest, however, are papers detailing solutions to practical problems. Those papers selected for presentation will appear in the Conference Proceedings, which are published by the Computer Society of the IEEE. HICSS-23 is sponsored by the University of Hawaii in cooperation with the ACM, the IEEE Computer Society, and the Pacific Research Institute for Information Systems and Management (PRIISM). Submissions are solicited in the areas: (1)The application of neural nets to model business tasks performed by people (e.g. Dutta and Shekhar paper on Applying Neural Nets to Rating Bonds, ICNN, 1988, Vol. II, pp. 443-450) (2)The development of neural nets to model human decision tasks (e.g. Gluck and Bower, Journal of Experimental Psychology: General, 117(3), 227-247) (3)The application of neural nets to improving modeling tools commonly used in business (e.g. neural networks to perform regression-like modeling) (4)The embedding of neural nets in commercial products (e.g. OCR scanners) Our order of preference is from (1) to (4) above. Papers which detail actual us age of neural networks are preferred to those which only propose uses. INSTRUCTIONS FOR SUBMITTING PAPERS: Manuscripts should be 12-26 typewritten, double-spaced pages in length. Do not send submissions that are significantly s horter or longer than this. Each manuscript will be subjected to refereeing. Manuscript papers should have a title pages that includes the title of the paper, full name(s) of its author(s), affiliation(s), complete mailing and electronic address(es), telephone number(s), and a 300- word abstract of the paper. DEADLINES A 300-word optional abstract may be submitted by April 30, 1989 by Email or mail. (If no reply to email in 7 days, send by U.S. mail also.) Feedback to author concerning abstract by May 31, 1989. Six paper copies of the manuscript are due by June 26, 1989. Notification of accepted papers by September 1, 1989. Accepted manuscripts, camera-ready, are due by October 1, 1989. SEND SUBMISSIONS AND QUESTIONS TO: Prof. William Remus College of Business Administration University of Hawaii 2404 Maile Way Honolulu, HI 96822 USA Tel.: (808)948-7608 EMAIL: CBADWRE@UHCCVM.BITNET FAX: (808)942-1591 ------------------------------ Subject: Connection between Hidden Markov Models and Connectionist Networks From: thanasis kehagias <ST401843%BROWNVM.BITNET@VMA.CC.CMU.EDU> Date: Mon, 13 Feb 89 00:47:00 -0500 The following paper explores the connection between Hidden Markov Models and Connectionist networks. anybody interested in a copy, email me. If you have a TeX setup I will send you the dvi file. else give me your physical mail address. OPTIMAL CONTROL FOR TRAINING THE MISSING LINK BETWEEN HIDDEN MARKOV MODELS AND CONNECTIONIST NETWORKS by Athanasios Kehagias Division of Applied Mathematics Brown University Providence, RI 02912 ABSTRACT For every Hidden Markov Modle there is a set of forward probabilities that need to be computed for both the recognition and training problem . These probabilties are computed recursively and hence the computation can be performed by a multistage , feedforward network that we will call Hidden Markov Model Network (HMMN). This network has exactly the same architecture as the standard Connectionist Network(CN). Furthermore, training a Hidden Markov Model is equivalent to optimizing a function of the HMMN; training a CN is equivalent to minimizing a function of the CN. Due to the multistage feedforward architecture, both problems can be seen as Optimal Control problems. By applying standard Optimal Control techniques, we discover in both problems that certain back propagating quantities (backward probabilities for HMMN, backward propogated errors for CN) are of crucial importance for the solution. So HMMN's and CN's are similar both in architecture and training. ************** i was influenced in this research by the work of H. Bourlard and C.C. Wellekens (the HMM- CN connection) and Y. leCun (Optimal Control applications in CN's). as I was finishing my paper I received a message by J.N. Hwang saying that he and S.Y. Kung have written a paper that includes similar results. Thanasis Kehagias ------------------------------ Subject: EURASIP Workshop on Neural Nets From: uunet!mcvax!inesc!alf!lba (Luis Borges de Almeida) Date: Mon, 06 Mar 89 16:37:30 +0000 EURASIP WORKSHOP ON NEURAL NETWORKS Sesimbra, Portugal February 15-17, 1990 ANNOUNCEMENT AND CALL FOR PAPERS The workshop will be held at the Hotel do Mar in Sesimbra, Portugal. It will take place in 1990, from February 15 morning to 17 noon, and will be sponsored by EURASIP, the European Association for Signal Processing. It will be open to participants from all countries, both from inside and outside of Europe. Contributions from all fields related to the neural network area are welcome. A (non-exclusive) list of topics is given below. Care is being taken to ensure that the workshop will have a high level of quality. Proposed contributions will be evaluated by an international technical committee. A proceedings volume will be published, and will be handed to participants at the beginning of the workshop. The number of participants will be limited to 50. Full contributions will take the form of oral presentations, and will correspond to papers in the proceedings. Some short contributions will also be accepted, for presentation of ongoing work, projects (ESPRIT, BRAIN, DARPA,...), etc. They will be presented in poster format, and will not originate any written publication. A small number of non-contributing participants may also be accepted. The official language of the workshop will be English. TOPICS: - signal processing (speech, image,...) - pattern recognition - algorithms (training procedures, new structures, speedups,...) - generalization - implementation - specific applications where NN have been proved better than other approaches - industrial projects and realizations Submissions, both for long and for short contributions, will consist of (strictly) 2-page summaries. Three copies should be sent directly to the Technical Chairman, at the address given below. The calendar for contributions is as follows: Full contributions Short contributions Deadline for submission June 1, 1989 Oct 1, 1989 Notif. of acceptance Sept 1, 1989 Nov 15, 1989 Camera-ready paper Nov 1, 1989 SESIMBRA... ... is a fishermens village, located in a nice region about 30 km south of Lisbon. Special transportation from/to Lisbon will be arranged. The workshop will end on a Saturday at lunch time; therefore, the participants will have the option of either flying back home in the afternoon, or staying for sightseeing for the remainder of the weekend in Sesimbra and/or Lisbon. An optional program for accompanying persons is being organized. For further information, send the coupon below to the general chairman, or contact directly. ORGANIZING COMMITTEE: GENERAL CHAIRMAN Luis B. Almeida INESC Apartado 10105 P-1017 LISBOA CODEX PORTUGAL Phone: +351-1-544607. Fax: +351-1-525843. E-mail: {any backbone, uunet}!mcvax!inesc!lba TECHNICAL CHAIRMAN Christian Wellekens Philips Research Laboratory Av. Van Becelaere 2 Box 8 B-1170 BRUSSELS BELGIUM Phone: +32-2-6742275 TECHNICAL COMMITTEE John Bridle (Royal Signal and Radar Establishment, Malvern, UK) Herve Bourlard (Intern. Computer Science Institute, Berkeley, USA) Frank Fallside (University of Cambridge, Cambridge, UK) Francoise Fogelman (Ecole de H. Etudes en Informatique, Paris, France) Jeanny Herault (Institut Nat. Polytech. de Grenoble, Grenoble, France) Larry Jackel (AT&T Bell Labs, Holmdel, NJ, USA) Renato de Mori (McGill University, Montreal, Canada) REGISTRATION, FINANCE, LOCAL ARRANGEMENTS Joao Bilhim INESC Apartado 10105 P-1017 LISBOA CODEX PORTUGAL Phone: +351-1-545150. Fax: +351-1-525843. WORKSHOP SPONSOR: EURASIP - European Association for Signal Processing CO-SPONSORS: INESC - Instituto de Engenharia de Sistemas e Computadores, Lisbon, Portugal IEEE, Portugal Section *-------------------------------- cut here ---------------------------------* Please keep me informed about the EURASIP Workshop on Neural Networks Name: University/Company: Address: Phone: E-mail: [ ] I plan to attend the workshop I plan to submit a contribution [ ] full [ ] short Preliminary title: (send to Luis B. Almeida, at address given above) ------------------------------ Subject: NIPS POST-MEETING WORKSHOPS From: Stephen J Hanson <jose@tractatus.bellcore.com> Date: Tue, 14 Feb 89 17:22:22 -0500 NIPS-89 POST-CONFERENCE WORKSHOPS DECEMBER 1-2, 1989 REQUEST FOR PROPOSALS Following the regular NIPS program, workshops on current topics in Neural Information Processing will be held on December 1 and 2, 1989, at a ski resort near Denver. Proposals by qualified individuals interested in chairing one of these workshops are solicited. Past topics have included: Rules and Connectionist Models; Speech, Neural Networks and Hidden Markov Models; Imaging Techniques in Neurobiology; Computational Complexity Issues; Fault Tolerance in Neural Networks; Benchmarking and Comparing Neural Network Applications; Architectural Issues; Fast Training Techniques. The format of the workshops is informal. Beyond reporting on past research, their goal is to provide a forum for scientists actively working in the field to freely discuss current issues of concern and interest. Sessions will meet in the morning and in the afternoon of both days, with free time in between for ongoing individual exchange or outdoor activities. Specific open and/or controversial issues are encouraged and preferred as workshop topics. Individuals interested in chairing a workshop must propose a topic of current interest and must be willing to accept responsibility for their group's discussion. Discussion leaders' responsibilities include: arrange brief informal presentations by experts working on this topic, moderate or lead the discussion; and report its high points, findings and conclusions to the group during evening plenary sessions. Submission Procedure: Interested parties should submit a short proposal for a workshop of interest by May 30, 1989. Proposals should include a title and a short description of what the workshop is to address and accomplish. It should state why the topic is of interest or controversial, why it should be discussed and what the targeted group of participants is. In addition, please send a brief resume of the prospective workshop chair, list of publications and evidence of scholarship in the field of interest. Mail submissions to: Kathie Hibbard NIPS89 Local Committee Engineering Center Campus Box 425 Boulder, CO, 80309-0425 Name, mailing address, phone number, and e-mail net address (if applicable) should be on all submissions. Workshop Organizing Committee: Alex Waibel, Carnegie-Mellon, Workshop Chairman; Howard Wachtel, University of Colorado, Workshop Local Arrangements; Kathie Hibbard, University of Colorado, NIPS General Local Arrangements; PROPOSALS MUST BE RECEIVED BY MAY 30, 1989. ------------------------------ Subject: Preprint - Performance of a Stochastic Learning Microchip From: Selma M Kaufman <smk@flash.bellcore.com> Date: Fri, 17 Feb 89 09:45:33 -0500 Performance of a Stochastic Learning Microchip Joshua Alspector, Bhusan Gupta, and Robert B. Allen We have fabricated a test chip in 2 micron CMOS that can perform supervised learning in a manner similar to the Boltzmann machine. Patterns can be presented to it at 100,000 per second. The chip learns to solve the XOR problem in a few milliseconds. We also have demonstrated the capability to do unsupervised competitive learning with it. The functions of the chip components are exam- ined and the performance is assessed. For copies contact: Selma Kaufman, smk@flash.bellcore.com ------------------------------ Subject: Rules and Variables in a Connectionist Reasoning System From: Lokendra Shastri <Shastri@cis.upenn.edu> Date: Sun, 26 Feb 89 20:58:00 -0500 Technical report announcement, please send requests to glenda@cis.upenn.edu A Connectionist System for Rule Based Reasoning with Multi-Place Predicates and Variables Lokendra Shastri and Venkat Ajjanagadde Computer and Information Science Department University of Pennsylvania Philadelphia, PA 19104 MS-CIS-8906 LINC LAB 141 Abstract McCarthy has observed that the representational power of most connectionist systems is restricted to unary predicates applied to a fixed object. More recently, Fodor and Pylyshyn have made a sweeping claim that connectionist systems cannot incorporate systematicity and compositionality. These comments suggest that representing structured knowledge in a connectionist network and using this knowledge in a systematic way is considered difficult if not impossible. The work reported in this paper demonstrates that a connectionist system can not only represent structured knowledge and display systematic behavior, but it can also do so with extreme efficiency. The paper describes a connectionist system that can represent knowledge expressed as rules and facts involving multi-place predicates (i.e., n-ary relations), and draw limited, but sound, inferences based on this knowledge. The system is extremely efficient - in fact, optimal, as it draws conclusions in time proportional to the length of the proof. It is observed that representing and reasoning with structured knowledge requires a solution to the variable binding problem. A solution to this problem using a multi-phase clock is proposed. The solution allows the system to maintain and propagate an arbitrary number of variable bindings during the reasoning process. The work also identifies constraints on the structure of inferential dependencies and the nature of quantification in individual rules that are required for efficient reasoning. These constraints may eventually help in modeling the remarkable human ability of performing certain inferences with extreme efficiency. ------------------------------ Subject: Talk at ICSI - DIFICIL From: collison%icsi.Berkeley.EDU@berkeley.edu (Alexandra Collison) Date: Wed, 15 Feb 89 13:58:02 -0800 The International Computer Science Institute is pleased to announce a talk: Dr. Susan Hollbach Weber University of Rochester Monday, February 27, 1989 at 2:30 p.m. "DIFICIL: Direct Inferences and Figurative Interpretation in a Connectionist Implementation of Language understanding." Given that conceptual categories possess properties (or slots) and values (or fillers), the structural relationships between these attributes can account for many complex behaviours, ranging from direct inferences to the interpretation of novel figures of speech. This talk presents a connectionist imple- mentation of a functional model of category structure in which categories are multi-faceted and each facet is functionally motivated. The resulting system, known as DIFICIL, captures a wide variety of cognitive effects. Direct inferences arise from literal adjective-noun combinations, where inferences are drawn about property values based on the named property value; for example, green apples are unripe and sour, and green grass is soft and cool. Property dominance effects indicate that the adjective `green' actually primes the property values `unripe' and `sour' for the category `apple'. Prototype effects arise within a given aspect of a category, as the tightly coupled property values interact with each other. Finally, the model provides a mechanism to interpret novel figurative adjective-noun combi- nations, such as `green idea': property abstraction hierarchies supply all possible interpretations suggested by the conceptual aspects normally associated with the adjective. This talk will be held in ICSI's Main Lecture Hall. 1947 Center Street, Suite 600, Berkeley, CA 94704 (On Center between Milvia and Martin Luther King Jr. Way) ------------------------------ Subject: Talk at ICSI - "Perceptual Organization for Computer Vision" From: collison%icsi.Berkeley.EDU@berkeley.edu (Alexandra Collison) Date: Wed, 22 Feb 89 12:35:43 -0800 The International Computer Science Institute is pleased to present a talk: Thursday, March 9, 1989 2:30 p.m. Rakesh Mohan Institute for Robotics and Intelligent Systems, University of Southern California "Perceptual Organization for Computer Vision" Our ability to detect structural relationships among similar image tokens is termed "perceptual organization". In this presen- tation, we will discuss the grouping of intensity edges into "col- lations" on the basis of the geometric relationships among them. These collations encode structural information which can aid various visual tasks such as object segmentation, correspondence processes (stereo, motion and model matching) and shape inference. We will present two vision systems based on perceptual organization, one to detect and describe buildings in aerial images and the other to segment 2D scenes. This talk will be held in the Main Lecture Hall at ICSI. 1947 Center Street, Suite 600, Berkeley, CA 94704 (On Center between Milvia and Martin Luther King Jr. Way) ------------------------------ Subject: Talk at ICSI - Spreading Activation Meets Back Propagation: From: collison%icsi.Berkeley.EDU@berkeley.edu (Alexandra Collison) Date: Fri, 17 Feb 89 15:44:46 -0800 The International Computer Science Institute is pleased to present a talk: Dr. James Hendler ICSI and University of Maryland, College Park Wednesday, February 22, 1989 12 noon Spreading Activation Meets Back Propagation: Towards higher level inferencing with distributed networks Connectionism has recently seen a major resurgence of interest among both artificial intelligence and cognitive science researchers. The spectrum of these neural network approaches is quite large, ranging from structured models, in which individual network units carry meaning, through distributed models of weighted networks with learning algorithms. Very encouraging results, particularly in ``low-level'' perceptual and signal processing tasks, are being reported across the entire spectrum of these models. These models have had more limited success, however, in those ``higher cognitive'' areas where symbolic models have traditionally shown promise: expert reasoning, planning, and natural language processing. However, although connectionist techniques have had only limited success in such cognitive tasks, for a system to provide both ``low-level'' perceptual functionality as well as demonstrating high-level cognitive abilities, it must be able to capture the best features of each of the competing paradigms. In this talk we discuss several steps towards providing such a system by examining various models of spreading activation inferencing on networks created by parallel distributed processing learning techniques. This talk will be held in the ICSI Main Lecture Hall. 1947 Center Street, Suite 600, Berkeley, CA 94704 (On Center between Milvia and Martin Luther King Jr. Way) ------------------------------ Subject: Talk at ICSI - The Sphinx Speech Recognition System From: collison%icsi.Berkeley.EDU@berkeley.edu (Alexandra Collison) Date: Fri, 24 Feb 89 12:12:00 -0800 The International Computer Science Institute is pleased to present a talk: Dr. Kai-Fu Lee Computer Science Department Carnegie Mellon University Pittsburgh, Pennsylvania "The Sphinx Speech Recognition System" In this talk, I will describe SPHINX, the first large-vocabulary speaker-independent continuous speech recognition system. First, an overview of the system will be presented. Next, I will describe some of our recent enhancements, including: - generalized triphone models - word duration modeling - function-phrase modeling - between-word coarticulation modeling - corrective training Our most recent results with the 997-word resource management task are: 96% word accuracy with a grammar (perplexity 60), and 82% without grammar (perplexity 997). I will also describe our recent results with: - Speech recognition without vocabulary-specific training. - Using neural networks for continuous speech recognition. This talk will be held in the Main Lecture Hall at ICSI. 1947 Center Street, Suite 600, Berkeley, CA 94704 (On Center between Milvia and Martin Luther King Jr. Way) ------------------------------ Subject: TR - ANNs and Sequential Paraphrasing of Script-Based Stories From: Risto Miikkulainen <risto@CS.UCLA.EDU> Organization: UCLA Computer Science Department Date: Thu, 23 Feb 89 14:15:22 -0800 [ Please send requests to valerie@cs.ucla.edu ] A Modular Neural Network Architecture for Sequential Paraphrasing of Script-Based Stories Risto Miikkulainen and Michael G. Dyer Artificial Intelligence Laboratory Computer Science Department University of California, Los Angeles, CA 90024 Abstract We have applied sequential recurrent neural networks to a fairly high-level cognitive task, i.e. paraphrasing script-based stories. Using hierarchically organized modular subnetworks, which are trained separately and in parallel, the complexity of the task is reduced by effectively dividing it into subgoals. The system uses sequential natural language input and output, and develops its own I/O representations for the words. The representations are stored in an external global lexicon, and they are adjusted in the course of training by all four subnetworks simultaneously, according to the FGREP-method. By concatenating a unique identification with the resulting representation, an arbitrary number of instances of the same word type can be created and used in the stories. The system is able to produce a fully expanded paraphrase of the story from only a few sentences, i.e. the unmentioned events are inferred. The word instances are correctly bound to their roles, and simple plausible inferences of the variable content of the story are made in the process. ------------------------------ Subject: TR - ANNs in Robot Motion Planning From: mel@cougar.ccsr.uiuc.edu (Bartlett Mel) Date: Thu, 09 Feb 89 12:26:34 -0600 The following thesis/TR is now available--about 50% of it is dedicated to relations to traditional methods in robotics, and to psychological and biological issues... MURPHY: A Neurally-Inspired Connectionist Approach to Learning and Performance in Vision-Based Robot Motion Planning Bartlett W. Mel Center for Complex Systems Research Beckman Institute, University of Illinois Many aspects of intelligent animal behavior require an understanding of the complex spatial relationships between the body and its parts and the coordinate systems of the external world. This thesis deals specifically with the problem of guiding a multi-link arm to a visual target in the presence of obstacles. A simple vision-based kinematic controller and motion planner based on a connectionist network architecture has been developed, called MURPHY. The physical setup consists of a video camera and a Rhino XR-3 robot arm with three joints that move in the image plane of the camera. We assume no a priori model of arm kinematics or of the imaging characteristics of the camera/visual system, and no sophisticated built-in algorithms for obstacle avoidance. Instead, MURPHY builds a model of his arm through a combination of physical and ``mental'' practice, and then uses simple heuristic search with mental images of his arm to solve visually-guided reaching problems in the presence of obstacles whose traditional algorithmic solutions are extremely complex. MURPHY differs from previous approaches to robot motion-planning primarily in his use of an explicit full-visual-field representation of the workspace. Several other aspects of MURPHY's design are unusual, including the sigma-pi synaptic learning rule, the teacherless training paradigm, and the integration of sequential control within an otherwise connectionist architecture. In concluding sections we outline a series of strong correspondences between the representations and algorithms used by MURPHY, and the psychology, physiology, and neural bases for the programming and control of directed, voluntary arm movements in humans and animals. You can write to me: mel@complex.ccsr.uiuc.edu, or judi jr@complex.ccsr.uiuc.edu. Out computers go down on Feb. 13 for 2 days, so if you want one then, call (217)244-4250 instead. -Bartlett Mel ------------------------------ Subject: TR - Dynamic Node Creating in Back-Prop Nets From: biafore@beowulf.ucsd.edu (Louis Steven Biafore) Organization: Computer Science & Engineering Dept. U.C. San Diego Date: Tue, 07 Mar 89 20:43:19 +0000 The following technical report is now available: DYNAMIC NODE CREATION IN BACKPROPAGATION NETWORKS Timur Ash ash@ucsd.edu Abstract Large backpropagation (BP) networks are very difficult to train. This fact complicates the process of iteratively testing different sized networks (i.e., networks with dif- ferent numbers of hidden layer units) to find one that pro- vides a good mapping approximation. This paper introduces a new method called Dynamic Node Creation (DNC) that attacks both of these issues (training large networks and testing networks with different numbers of hidden layer units). DNC sequentially adds nodes one at a time to the hidden layer(s) of the network until the desired approximation accuracy is achieved. Simulation results for parity, symmetry, binary addition, and the encoder problem are presented. The pro- cedure was capable of finding known minimal topologies in many cases, and was always within three nodes of the minimum. Computational expense for finding the solutions was comparable to training normal BP networks with the same final topologies. Starting out with fewer nodes than needed to solve the problem actually seems to help find a solution. The method yielded a solution for every problem tried. BP applied to the same large networks with randomized initial weights was unable, after repeated attempts, to replicate some minimum solutions found by DNC. Requests for reprints should be sent to the Institute for Cognitive Science, C-015; University of California, San Diego; La Jolla, CA 92093. (ICS Report 8901) ------------------------------ Subject: TR - Learning State Space Trajectories in Recurrent ANNs From: Barak.Pearlmutter@F.GP.CS.CMU.EDU Date: 16 Feb 89 19:33:00 -0500 The following tech report is available. It is a substantially expanded version of a paper of the same title that appeared in the proceedings of the 1988 CMU Connectionist Models Summer School. Learning State Space Trajectories in Recurrent Neural Networks Barak A. Pearlmutter ABSTRACT We describe a number of procedures for finding $\partial E/\partial w_{ij}$ where $E$ is an error functional of the temporal trajectory of the states of a continuous recurrent network and $w_{ij}$ are the weights of that network. Computing these quantities allows one to perform gradient descent in the weights to minimize $E$, so these procedures form the kernels of connectionist learning algorithms. Simulations in which networks are taught to move through limit cycles are shown. We also describe a number of elaborations of the basic idea, such as mutable time delays and teacher forcing, and conclude with a complexity analysis. This type of network seems particularly suited for temporally continuous domains, such as signal processing, control, and speech. Overseas copies are sent first class so there is no need to make special arrangements for rapid delivery. Requests for copies should be sent to Catherine Copetas School of Computer Science Carnegie Mellon University Pittsburgh, PA 15213 or Copetas@CS.CMU.EDU by computer mail. Ask for CMU-CS-88-191. ------------------------------ Subject: TR - Speeding up ANNs in the "Real World" From: Josiah Hoskins <joho%sw.MCC.COM@MCC.COM> Date: Thu, 02 Mar 89 12:18:40 -0600 The following tech report is available. Speeding Up Artificial Neural Networks in the "Real" World Josiah C. Hoskins A new heuristic, called focused-attention backpropagation (FAB) learning, is introduced. FAB enhances the backpropagation pro- cedure by focusing attention on the exemplar patterns that are most difficult to learn. Results are reported using FAB learning to train multilayer feed-forward artificial neural networks to represent real-valued elementary functions. The rate of learning observed using FAB is 1.5 to 10 times faster than backpropagation. Request for copies should refer to MCC Technical Report Number STP-049-89 and should be sent to Kintner@mcc.com or to Josiah C. Hoskins MCC - Software Technology Program AT&T: (512) 338-3684 9390 Research Blvd, Kaleido II Bldg. UUCP/USENET: milano!joho Austin, Texas 78759 ARPA/INTERNET: joho@mcc.com ------------------------------ End of Neurons Digest *********************