COE%PLU@ames-io.ARPA (02/02/88)
*********************** OPEN TECHNICAL MEETING *********************** IEEE Computer Society, Santa Clara Valley Chapter Tuesday Febuary 9, 1988 8:00 p.m. Hewlett-Packard Cupertino (Wolfe & Homestead) Building 48, Oak Room "CAPACITY FOR PATTERNS AND SEQUENCES IN KANERVA'S SDM AS COMPARED TO OTHER ASSOCIATIVE MEMORY MODELS" ABSTRACT: Dr. James Keeler of Stanford University will be speaking on the information capacity of Kanerva's Sparse, Distributed Memory (SDM) and Hopfield-type neural networks will be discussed. Using certain approximations, it is shown that the total information stored in these systems is proportional to the number connections in the network. The proportionality constant is the same for the SDM and Hopfield-type models independent of the particular model, or the order of the model. This same analysis can be used to show that the SDM can store sequences of spatiotemporal patterns, and the addition of time-delayed connections allows the retrieval of context dependent temporal patterns with varying time delays. Dr. Keeler Received his Ph.D in Physics from U. C. San Diego, March 1987. His dissertation was on reaction-diffusion systems and neural network models. He is now a Postdoctoral student at Stanford University's Department of Chemistry working on neural network models, as well as consulting for Penti Kanerva's (RIACS, NASA Ames Research Center) SDM research group. For additional information contact Coe Miles-Schlichting: coe@pluto.arc.nasa.gov or (408) 279-4773