neuron-request@HPLABS.HP.COM ("Neuron-Digest Moderator Peter Marvit") (02/04/89)
Neuron Digest Friday, 3 Feb 1989 Volume 5 : Issue 7 Today's Topics: Joint Conference on Neural Networks new tech report A WORKSHOP ON CELLULAR AUTOMATA Genetic Algorithms and Neural Networks Tech. Report Available oh boy, more tech reports... Technical Report: LAIR 89-JP-NIPS Kolmogorov's superposition theorem 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: Joint Conference on Neural Networks From: lehr@isl.stanford.EDU (Michael Lehr) Organization: Stanford University Date: 06 Jan 89 01:27:31 +0000 [[ Editor's note: The deadline has actually been advanced to 15 February! -PM]] An announcement that Bernie Widrow asked me to submit to the digest: December 29, 1988 Fellow INNS Members: In accord with the decisions of the INNS Board of Governors and its Executive Committee, a negotiation with the IEEE has been undertaken on the subject of joint technical meetings. The INNS and the IEEE have succeeded in the negotiation! The first joint meeting will be held June 18-22, 1989 at the Sheraton Washington Hotel. Please see the attached announcement and call for papers. In the past, the IEEE has held a series of meetings called ICNN (International Conference on Neural Networks) and we held our first annual meeting in Boston and called it the INNS Conference. In the future these conferences will be combined and called the IJCNN, International Joint Conference on Neural Networks. The IEEE was scheduled to have the next ICNN in June 1989 in Washington D.C. This will superseded by the first joint meeting, IJCNN-89. We were scheduled to have our next INNS meeting at the Omni Shoreham Hotel in Washington, DC in September 1989. The conference date has been changed to January 1990 and this will become the second joint meeting, IJCNN-W'90 (IJCNN-Winter 1990). A call for papers will go out in about two months. The need for two international conferences per year in the United States was clear. During the past summer, both the IEEE San Diego meeting and the INNS Boston meeting were highly successful. Although their technical programs were very similar and they were spaced only six weeks apart, both meetings were attended by about 1700 people. There is a need for these U.S. meetings, and geographical diversity is the key. We are also planning to have meetings in Europe and Asia. The following is a partial list of reasons expressed by various INNS members on why joint meetings with IEEE are desired: o Engender a spirit of cooperation among scientists and engineers in the field. o Bring together life scientists with technologists for the betterment of all areas of neural network research. o Enhance the scope and size of the field, facilitating public and private support for research and commercialization efforts. o Increase the ability of the neural network industry to exhibit their technical developments and products. o Avoid scheduling conflicts. Our agreement with the IEEE calls for a review after the first two meetings. If we are successful, then we go on together for two more meetings. And so forth. Once we get over the initial transients, we will try to have winter meetings on the West Coast and summer meetings on the East Coast. The two societies will alternate in taking primary responsibility for meeting organization. The agreement is symmetrical, and both societies are 50-50 partners in the meetings. The IJCNN conferences are for the benefit of both IEEE and INNS members. Both memberships will enjoy reduced conference fees at all joint meetings, with students attending at even more reduced rates. Fellow INNS'ers, I urge you to consider the attached call for papers! IJCNN-89 serves as our annual meeting for 1989. The deadline on the call for papers shows February 1, 1989, but has been extended to February 15, 1989 for INNS members. This is the absolute cut-off date. Get your papers into Nomi Feldman, and incidentally don't forget to renew your INNS membership. My best wishes to you for the New Year, Dr. Bernard Widrow, President INNS - -------------------------call for papers------------------------- N E U R A L N E T W O R K S CALL FOR PAPERS International Joint Conference on Neural Networks June 18-22, 1989 Washington, D.C. The 1989 IEEE/INNS International Joint Conference on Neural Networks (IJCNN-89) will be held at the Sheraton Washington Hotel in Washington, D.C., USA from June 18-22, 1989. IJCNN-89 is the first conference in a new series devoted to the technology and science of neurocomputing and neural networks in all of their aspects. The series replaces the previous IEEE ICNN and INNS Annual Meeting series and is jointly sponsored by the IEEE Technical Activities Board Neural Network Committee and the International Neural Network Society (INNS). IJCNN-89 will be the only major neural network meeting of 1989 (IEEE ICNN-89 and the 1989 INNS Annual Meeting have both been cancelled). Thus, it behoves all members of the neural network community who have important new results to present to prepare their papers now and submit them by the IJCNN-89 deadline of 1 FEBRUARY 1989. To provide maximum value to participants, the full text of those papers presented orally in the technical sessions will be published in the Conference Proceedings (along with some particularly outstanding papers from the Poster Sessions). The title, author name, author affiliation, and abstract portions of all Poster Session papers not published in full will also be published in the Proceedings. The Conference Proceedings will be distributed *at the registration desk* to all regular conference registrants as well as to all student registrants. The conference will include a day of tutorials (June 18), the exhibit hall (the neurocomputing industry's primary annual trade show), plenary talks, and social events. Mark your calendar today and plan to attend IJCNN-89--the definitive annual progress report on the neurocomputing revolution! **DEADLINE for submission of papers for IJCNN-89 is FEBRUARY 1, 1989** [[Editor's note: I understand this has been extended to 15 Feb 89. -PM]] Papers of 8 pages or less are solicited in the following areas: * Real World Applications * Neural Network Architectures and Theory * Supervised Learning Theory * Reinforcement Learning Theory * Robotics and Control * Optical Neurocomputers * Optimization * Associative Memory * Image Analysis * Self-Organization * Neurobiological Models * Vision * Electronic Neurocomputers Papers should be prepared in standard IEEE Conference Proceedings Format, and typed on the special forms provided in the Author's Kit. Indicate which of the above subject areas you wish your paper included in and whether you wish your paper to be considered for oral presentation at a technical session, presentation as a poster at a poster session, or both. Papers submitted for oral presentation may, at the referees' discretion, be designated for poster presentation instead, if they feel this would be more appropriate. FULL PAPERS in camera-ready form (1 original on Author's Kit forms and 5 reduced 8.5" x 11" copies) should be submitted to Nomi Feldman, Conference Coordinator, at the address below. For more details, or to request your IEEE Author's Kit, call or write: Nomi Feldman, IJCNN-89 Conference Coordinator 3770 Tansy Street, San Diego, CA 92121 (619) 453-6222 ------------------------------ Subject: new tech report From: Geoffrey Hinton <hinton@ai.toronto.edu> Date: Tue, 10 Jan 89 10:09:11 -0500 The following report can be obtained by sending an email request to carol@ai.toronto.edu If this fails try carol%ai.toronto.edu@relay.cs.net Please do not send email to me about it. "Deterministic Boltzmann Learning Performs Steepest Descent in Weight-space." Geoffrey E. Hinton Department of Computer Science University of Toronto Technical report CRG-TR-89-1 ABSTRACT The Boltzmann machine learning procedure has been successfully applied in deterministic networks of analog units that use a mean field approximation to efficiently simulate a truly stochastic system {Peterson and Anderson, 1987}. This type of ``deterministic Boltzmann machine'' (DBM) learns much faster than the equivalent ``stochastic Boltzmann machine'' (SBM), but since the learning procedure for DBM's is only based on an analogy with SBM's, there is no existing proof that it performs gradient descent in any function, and it has only been justified by simulations. By using the appropriate interpretation for the way in which a DBM represents the probability of an output vector given an input vector, it is shown that the DBM performs steepest descent in the same function as the original SBM, except at rare discontinuities. A very simple way of forcing the weights to become symmetrical is also described, and this makes the DBM more biologically plausible than back-propagation. ------------------------------ Subject: A WORKSHOP ON CELLULAR AUTOMATA From: Gerard Vichniac (617)873-2762 <gerard@bbn.com> Date: Sun, 15 Jan 89 23:56:23 -0500 A workshop on: ************************************************************ Cellular Automata and Modeling of Complex Physical Systems ************************************************************ will be held in les Houches, near Chamonix, France, from February 21 to March 2, 1989. The organizers are Roger Bidaux (Saclay), Paul Manneville (Saclay), Yves Pomeau (Ecole Normale, Paris), and Gerard Vichniac (MIT and BBN). The topics will include: - - automata and discrete dynamical systems, - - lattice-gas automata for fluid dynamics, - - applications to solid-state physics (in particular, models of growth and pattern formation), - - parallel computation in statistical mechanics (in particular, in the Ising model), - - dedicated cellular-automata machines. Workshops at les Houches are traditionally informal, there will be about five talks a day, and ample time will be left for discussion. A fee of 3700 FF includes full board lodging at the Physics Center in les Houches. Contact: Paul Manneville, Roger Bidaux or Gerard Vichniac Bitnet: MANNEV @ FRSAC11 Internet: gerard@alexander.bbn.com tel.: 33 - 1 69 08 75 35 tel.: (617) 253 5893 (MIT) fax: 33 - 1 69 08 81 20 (617) 873 2762 (BBN) telex: 604641 ENERG F BITNET: MANNEV @ FRSAC11 ------------------------------ Subject: Genetic Algorithms and Neural Networks From: DMONTANA%COOPER@rcca.bbn.com Date: Mon, 16 Jan 89 20:53:00 -0500 We have written a paper that may be of interest to a number of readers on both of these mailing lists: "Training Feedforward Neural Networks Using Genetic Algorithms" David J. Montana and Lawrence Davis BBN Systems and Technologies Corp. 70 Fawcett St. Cambridge, MA 02138 ABSTRACT Multilayered feedforward neural networks possess a number of properties which make them particularly suited to complex pattern classification problems. However, their application to some real-world problems has been hampered by the lack of a training algorithm which reliably finds a nearly globally optimal set of weights in a relatively short time. Genetic algorithms are a class of optimization procedures which are good at exploring a large and complex space in an intelligent way to find values close to the global optimum. Hence, they are well suited to the problem of training feedforward networks. In this paper, we describe a set of experiments performed on a set of data from a sonar image classification problem. These experiments both 1) illustrate the improvements gained by using a genetic algorithm rather than backpropagation and 2) chronicle the evolution of the performance of the genetic algorithm as we added more and more domain-specific knowledge into it. In addition to outperforming backpropagation on the network we were investigating (which had nodes with sigmoidal transfer function), the genetic algorithm has the advantage of not requiring the nodes to have differentiable transfer functions. In particular, it can be applied to a network with thresholded nodes. Requests for copies of the paper should be addressed to dmontana@bbn.com. Thanks for considering this submission. Dave Montana ------------------------------ Subject: Tech. Report Available From: <MUMME%IDCSVAX.BITNET@CUNYVM.CUNY.EDU> Date: Tue, 17 Jan 89 20:22:00 -0800 The following tech. report is available from the University of Illinois Dept. of Computer Science: UIUCDCS-R-88-1485 STORAGE CAPACITY OF THE LINEAR ASSOCIATOR: BEGINNINGS OF A THEORY OF COMPUTATIONAL MEMORY by Dean C. Mumme May, 1988 ABSTRACT This thesis presents a characterization of a simple connectionist-system, the linear-associator, as both a memory and a classifier. Toward this end, a theory of memory based on information-theory is devised. The principles of the information-theory of memory are then used in conjunction with the dynamics of the linear-associator to discern its storage capacity and classification capabilities as they scale with system size. To determine storage capacity, a set of M vector-pairs called "items" are stored in an associator with N connection-weights. The number of bits of information stored by the system is then determined to be about (N/2)logM. The maximum number of items storable is found to be half the number of weights so that the information capacity of the system is quantified to be (N/2)logN. Classification capability is determined by allowing vectors not stored by the associator to appear its input. Conditions necessary for the associator to make a correct response are derived from constraints of information theory and the geometry of the space of input-vectors. Results include derivation of the information-throughput of the associator, the amount of information that that must be present in an input-vector and the number of vectors that can be classified by an associator of a given size with a given storage load. Figures of merit are obtained that allow comparison of capabilities of general memory/classifier systems. For an associator with a simple non-linarity on its output, the merit figures are evaluated and shown to be suboptimal. Constant attention is devoted to relative parameter size required to obtain the derived performance characteristics. Large systems are shown to perform nearest the optimum performance limits and suggestions are made concerning system architecture needed for best results. Finally, avenues for extension of the theory to more general systems are indicated. This tech. report is essentially my Ph.D. thesis completed last May and can be obtained by sending e-mail to: erna@a.cs.uiuc.edu Please do not send requests to me since I now live in Idaho and don't have access to the tech. reports. Comments, questions and suggestions about the work can be sent directly to me at the address below. Thank You! Dean C. Mumme bitnet: mumme@idcsvax Dept. of Computer Science University of Idaho Moscow, ID 83843 ------------------------------ Subject: oh boy, more tech reports... From: Michael C. Mozer <mozer%neuron@boulder.Colorado.EDU> Date: Wed, 18 Jan 89 14:19:46 -0700 Please e-mail requests to "kate@boulder.colorado.edu". Skeletonization: A Technique for Trimming the Fat from a Network via Relevance Assessment Michael C. Mozer Paul Smolensky University of Colorado Department of Computer Science Tech Report # CU-CS-421-89 This paper proposes a means of using the knowledge in a network to deter- mine the functionality or _relevance_ of individual units, both for the purpose of understanding the network's behavior and improving its perfor- mance. The basic idea is to iteratively train the network to a certain performance criterion, compute a measure of relevance that identifies which input or hidden units are most critical to performance, and automatically trim the least relevant units. This _skeletonization_ technique can be used to simplify networks by eliminating units that convey redundant infor- mation; to improve learning performance by first learning with spare hidden units and then trimming the unnecessary ones away, thereby constraining generalization; and to understand the behavior of networks in terms of minimal "rules." [An abridged version of this TR will appear in NIPS proceedings.] - --------------------------------------------------------------------------- And while I'm at it, some other recent junk, I mean stuff... A Focused Back-Propagation Algorithm for Temporal Pattern Recognition Michael C. Mozer University of Toronto Connectionist Research Group Tech Report # CRG-TR-88-3 Time is at the heart of many pattern recognition tasks, e.g., speech recog- nition. However, connectionist learning algorithms to date are not well- suited for dealing with time-varying input patterns. This paper introduces a specialized connectionist architecture and corresponding specialization of the back-propagation learning algorithm that operates efficiently on temporal sequences. The key feature of the architecture is a layer of self-connected hidden units that integrate their current value with the new input at each time step to construct a static representation of the tem- poral input sequence. This architecture avoids two deficiencies found in other models of sequence recognition: first, it reduces the difficulty of temporal credit assignment by focusing the back propagated error signal; second, it eliminates the need for a buffer to hold the input sequence and/or intermediate activity levels. The latter property is due to the fact that during the forward (activation) phase, incremental activity _traces_ can be locally computed that hold all information necessary for back propagation in time. It is argued that this architecture should scale better than conventional recurrent architectures with respect to sequence length. The architecture has been used to implement a temporal version of Rumelhart and McClelland's verb past-tense model. The hidden units learn to behave something like Rumelhart and McClelland's "Wickelphones," a rich and flexible representation of temporal information. - --------------------------------------------------------------------------- A Connectionist Model of Selective Attention in Visual Perception Michael C. Mozer University of Toronto Connectionist Research Group Tech Report # CRG-TR-88-4 This paper describes a model of selective attention that is part of a con- nectionist object recognition system called MORSEL. MORSEL is capable of identifying multiple objects presented simultaneously on its "retina," but because of capacity limitations, MORSEL requires attention to prevent it from trying to do too much at once. Attentional selection is performed by a network of simple computing units that constructs a variable-diameter "spotlight" on the retina, allowing sensory information within the spotlight to be preferentially processed. Simulations of the model demon- strate that attention is more critical for less familiar items and that at- tention can be used to reduce inter-item crosstalk. The model suggests four distinct roles of attention in visual information processing, as well as a novel view of attentional selection that has characteristics of both early and late selection theories. ------------------------------ Subject: Technical Report: LAIR 89-JP-NIPS From: Jordan B. Pollack <pollack@cis.ohio-state.edu> Date: Fri, 20 Jan 89 15:40:09 -0500 Preprint of a NIPS paper is now available. Request LAIR 89-JP-NIPS From: Randy Miller CIS Dept/Ohio State University 2036 Neil Ave Columbus, OH 43210 or respond to this message. - ------------------------------------------------------------------------------ IMPLICATIONS OF RECURSIVE DISTRIBUTED REPRESENTATIONS Jordan B. Pollack Laboratory for AI Research Ohio State University Columbus, OH 43210 I will describe my recent results on the automatic development of fixed-width recursive distributed representations of variable-sized hierarchal data structures. One implication of this work is that certain types of AI-style data-structures can now be represented in fixed-width analog vectors. Simple inferences can be performed using the type of pattern associations that neural networks excel at. Another implication arises from noting that these representations become self-similar in the limit. Once this door to chaos is opened, many interesting new questions about the representational basis of intelligence emerge, and can (and will) be discussed. ------------------------------ Subject: Kolmogorov's superposition theorem From: sontag@fermat.rutgers.edu Date: Tue, 17 Jan 89 14:08:03 -0500 *** I am posting this for Professor Rui de Figuereido, a researcher in Control Theory and Circuits who does not subscribe to this list. Please direct cc's of all responses to his e-mail address (see below). -eduardo s. *** KOLMOGOROV'S SUPERPOSITION THEOREM AND ARTIFICIAL NEURAL NETWORKS Rui J. P. de Figueiredo Dept. of Electrical and Computer Engineering Rice University, Houston, TX 77251-1892 e-mail: rui@zeta.rice.edu The implementation of the Kolmogorov-Arnold-Sprecher Superposition Theorem [1-3] in terms of artificial neural networks was first presented and fully discussed by me in 1980 [4]. I also discussed, then [4], applications of these structures to statistical pattern recognition and image and multi- dimensional signal processing. However, I did not use the words "neural networks" in defining the underlying networks. For this reason, the current researchers on neural nets including Robert Hecht-Nielsen [5] do not seem to be aware of my contribution [4]. I hope that this note will help correct history. Incidentally, there is a misprint in [4]. In [4], please insert "no" in the statement before eqn.(4). That statement should read: "Sprecher showed that lambda can be any nonzero number which satisfies no equation ..." [1] A.K.Kolmogorov, "On the representation of continuous functions of several variables by superposition of continuous functions of one variable and addition," Dokl.Akad.Nauk.SSSR,Vol.114,pp.369-373,1957. [2] V.I.Arnol'd, "On functions of three variables," Dokl.Akad.Nauk.SSSR, Vol.114,pp.953-956,1957. [3] D.A.Sprecher, "An improvement in the superposition theorem of Kolmogorov," J.Math.Anal.Appl.,Vol.38,pp.208-213,1972. [4] Rui J.P.de Figueiredo, "Implications and applications of Kolmogorov's superposition theorem,"IEEE Trans.Auto.Contr.,Vol.AC-25,pp.1227-1231,1980. [5] R.Hecht-Nielsen, "Kolmogorov's mapping neural network existence theorem," IEEE 1st Int.Conf.on Neural Networks, San Diego,CA,June 21-24,1987,paper III-11. ------------------------------ End of Neurons Digest *********************