NEURON-Request@ti-csl.csc.ti.COM (NEURON-Digest moderator Michael Gately) (02/22/88)
NEURON Digest Sun Feb 21 18:38:24 CST 1988 Volume 3 / Issue 6 Today's Topics: Neural Network Simulator on Atari ST..?? Boston University Graduate Program Help explain the concepts of the future Seminar: The One Thousand Node Mieko NN Engine (UTA) Short Course: OCCAM at UTA Workshop - 1st Joint Tech Wkshp on NNs and Fuzzy Logic (Houston) MBL summer course NEURAL NETS Abstract for the NEURON DIGEST ---------------------------------------------------------------------- Date: 16 Feb 88 22:48:43 GMT From: Aspen2 - VAXstation II <aspen2.dec.com!boiko@decwrl.dec.com> Subject: Neural Network Simulator on Atari ST..?? Does anyone know if there is a NN simulator on the Atari ST..? I believe that this would be a very cost effective way to explore NN on a low cost and fairly quick personal system. I know that there are NN simulators available for the Mac and IBM...but what about the ST..? Any info would be helpful... Thanks -mike- Phone #: (603) 894-2659 ------------------------------ Date: Fri, 12 Feb 88 00:45:10 EST From: Michael Cohen <mike%bucasb.bu.edu@bu-it.bu.edu> Subject: Boston University Graduate Program BOSTON UNIVERSITY M.A. AND PH.D. PROGRAM IN COGNITIVE AND NEURAL SYSTEMS Boston University will offer an M.A. and Ph.D. program in Cognitive and Neural Systems starting in September, 1988, pending final approval. This program will present an integrated curriculum offering the full range of psychological, neurobiological, and computational concepts, models, and methods in the broad field variously called neural networks, connectionism, parallel distributed processing, and biological information processing, in which Boston University is an acknowledged leader. Each student will also be required to take an equal number of carefully selected courses in one or more core departments, such as psychology, biology, computer science, mathematics, or engineering. A limited number of full-time graduate research fellowships are expected to be available. For application materials, write Prof. Stephen Grossberg, Chairman CNS Program Center for Adaptive Systems Boston University 111 Cummington Street Boston, MA 02215 USA For further information, call (617) 353-7857 and ask for CNS information. ------------------------------ Subject: Help explain the concepts of the future Date: Fri, 12 Feb 88 07:38:08 -0500 From: G B Reilly <reilly@louie.udel.edu> The Franklin Institute Science Museum* will be opening the Futures Center in 1990. This is not a copy of EPCOT Center or a futuristic living room. It is exhibits to explain the new concepts in science and technology that will affect people's lives in the coming years. One section explains the concepts of robotics, computing, and artificial intelligence. We are interested in hearing what you believe the public needs to know about these areas and how they will affect their life in the next decade. Thank you for your cooperation. Brendan Reilly Curator ---- * The Franklin Institute is one of the oldest science museums in the country and has hands-on exhibits explaining science and technology which are visited by over one million people annually. ------------------------------ Date: Mon, 15 Feb 88 17:54:36 CST From: gately@resbld.csc.ti.com (Michael T. Gately) Subject: Seminar: The One Thousand Node Mieko NN Engine (UTA) The Automation & Robotics Research Institute cordially invites you to CORPORATE DAY Date: 26 February, 1988 Time: 2-5pm. Place: The Automation & Robotics Research Institute (affiliated with Universit of Texas at Arlington) 7300 Jack Newell Blvd South Fort Worth, Texas The Automation & Robotics Research Institute was established to coordinage the technological research of the College of Engineering of The University of Texas at Arlington with the needs of business and industry. Corporate Day's featured speaker will be Dr. D. J. Wallace, F.R.S., Tait Professor of Mathematical Physics of The University of Edinburgh, addressing the burgeoning field of artificial intelligence, in the Institute auditorium at 3:15pm. ------------------------------ Date: Mon, 15 Feb 88 18:09:40 CST From: gately@resbld.csc.ti.com (Michael T. Gately) Subject: Short Course: OCCAM at UTA OCCAM SHORT COURSE 24 February, 1988 @1:00pm to 26 February, 1988 @ 12:00noon Location: The Automation and Robotics Research Institute 7300 jack Newell Blvd South Fort Worth, TX 76118 Cost: $350 per person. Subsequent offerings will cost ~$600 per person. For further information: Tom Rethard Manager of Computer Systems The Automation and Robotics Institute The University of Texas at Arlington (817) 284-6103 ------------------------------ Date: Wed, 17 Feb 88 09:00:35 CST From: gately@resbld.csc.ti.com (Michael T. Gately) Subject: Workshop - 1st Joint Tech Wkshp on NNs and Fuzzy Logic (Houston) ANNOUNCEMENT 1988 First Joint Technology Workshop on Neural Networks and Fuzzy Logic Lyndon B. Johnson Space Center Gilruth Recreation Center Houston, Texas 2-3 May 1988 Co-sponsored by NASA - Johnson Space Center and the University of Houston - Clear Lake Nobel laureate Leon Cooper will address the workshop as keynote speaker discussion Artificial Neural Systems and their impact on technology development. Topics to be discussed by the foremost experts in their respective fields include: o Learning Algorithms o Network Architectures o Vision o Robotics o Neurobiological Connections o Speech Recognition and Synthesis o Fuzzy Set Theory and Applications o Control and Dynamics Processing o Space Applications o Fuzzy Logic and Neural Network Computers For further information or registration, please contact: Sandy Griffin AI Section - FM7 NASA - Johnson Space Center Houston, TX 77058 (713) 483-8071 ------------------------------ Date: Wed, 17 Feb 88 09:04:36 CST Subject: MBL summer course Announcement summer course in: Methods in Computational Neuroscience at The Marine Biological Laboratory Woods Hole, Massachusetts August 14 - September 3 Description: This three week course will survey basic computer modeling techniques used to study the nervous system at both single cell and neural network levels. Emphasis will be placed on using these models to explore the information processing capabilities of neural circuits. The principle aim of the course is to provide participants with the tools to simulate the functional properties of neural systems of interest to them and to understand the general advantages and pitfalls of this approach to understanding the nervous system. The first section of the course will focus on simulating the electrical properties of single neurons (compartmental models, active currents, interactions between synapses, calcium buffering and diffusion). The second part of the course will consider the numerical and graphical techniques necessary for modeling simple and complex networks of neurons. Examples of such simulations will be drawn from the invertebrate and vertebrate literature (visual system of the fly, swimming in Tritonia, learning in Hermissenda, mammalian olfactory and visual cortices). In the final section of the course, connectionist networks potentially relevant to perception and learning in real neural networks will be discussed and analyzed. Throughout the course, the actual machinery involved in modeling neural networks, including parallel computers like the Hypercube and Connection Machine will also be discussed. Between more formal lectures on these subjects, students will implement and test their own simulations in the computer laboratory. State-of-the-art single-user color work stations (6 Sun3-260s) will be provided for this purpose. These computers will run the general-purpose neural network simulating software that has been developed at Caltech, reducing the need for fundamental code writing. Students: The course is designed for advanced graduate students and postdoctoral fellows in biology, computer science, electrical engineering, physics, or psychology with an interest in computational neuroscience. Because of the heavy computer orientation of the Lab section of the course, a good computer background is highly desirable (preferably UNIX, C or PASCAL). In addition, students are expected to have a basic background in neurobiology. Course enrollment is limited to 12 students. Course Directors: James M. Bower and Christof Koch, Divisions of Biology and Engineering and Applied Science, Caltech. Course Faculty: Paul Adams (Stony Brook), Peter Getting (Iowa), John Hopfield (Caltech) Rudolfo Llinas (NYU), Richard Andersen (MIT), Idan Segev (Jerusalem), Terry Sejnowski (Johns Hopkins), David Van Essen (Caltech). Application: Requests for applications should be sent to: Office of Instruction Marine Biological Laboratory Woods Hole, Massachusetts 02543 Application Deadline: June 1, 1988 Board and Tuition: $1000 (Some fellowships are available) ------------------------------ Date: Wed 17 Feb 88 15:31:53-PDT From: HOSEIN@pluto.arc.nasa.gov Subject: NEURAL NETS Marc P. Hosein Intelligent Systems Technology Branch NASA Ames Research Center Mail Stop 244-4 Moffett Field, CA. 94035 (415) 694-6526 TO: Neural Network and Connectionist Researchers I am a research scientist in the Intelligent Systems Technology Branch of the Information Sciences division of NASA Ames Research Center. I am currently working on the Spaceborne VHSIC Multiprocessor System (SVMS) project under Dr. Henry Lum. In organizing a poster session on Neural Networks for the 1988 SOAR conference, I am gathering information on the current state of the field, as well as various technical and non-technical papers for distribution at the conference. The SOAR (Space Operations Automation and Robotics) workshop in automation and robotics is sponsored by NASA in conjunction with the USAF. The main objectives of the workshop are: 1) To establish communications between individuals and organizations involved in similar research and technology 2) To bring together project/program managers in open exchange through presentation of technical papers and panel discussions 3) To document in the proceedings a snapshot of USAF/NASA efforts in automation and robotics If you have papers or information to be included in a summary of the neural networks field, please mail them to me at the above address or on the arpanet. Even more importantly, I am looking for papers on research done or currently being done to incorporate as supplemental information and distribution material at the conference and beyond. Please feel free to call me at (415) 694-6526 or send mail on the arpanet to HOSEIN@AMES-PLUTO.ARPA if you have specific questions about the poster session or the conference. Thank you, Marc P. Hosein ------------------------------ Date: Mon, 15 Feb 88 14:10:20 EST From: jahir@athena.mit.edu Subject: Abstract for the NEURON DIGEST "A Methodology to Select Appropriate Learning Rate Parameters In Feed-Forward Networks" Jahir Pabon Graduate Research Assistant David Gossard Associate Professor Mechanical Engineering Department Massachusetts Institute of Technology Cambridge, MA 02139 Abstract The generalized delta rule was proposed by Rumelhart to modify the weights of the connections between units in adaptive feed-forward networks. This article presents a method to select appropriate values for the learning rate parameter used in the generalized delta rule. The validity of the method was evaluated through computer simulations of an adaptive network to control the orientation of an artificial eye. Several encoding parameters and output activation functions were tried. The optimal learning parameter value in each case was in good agreement with the value predicted by our method. Another important observation from the computer simulations was that a simple linear output activation function performed better than nonlinear ones. ----- Request for copies of this paper should include name and land address. Please send requests to JAHIR@ATHENA.MIT.EDU ------------------------------ End of NEURON-Digest ********************