Accession Number : ADA297003

Title :   Adaptive Optical Radial Basis Function Neural Network Classifier.

Descriptive Note : Rept. for Oct 93-Dec 94,

Corporate Author : ROME LAB GRIFFISS AFB NY

Personal Author(s) : Foor, Wesley E.

PDF Url : ADA297003

Report Date : DEC 1994

Pagination or Media Count : 42

Abstract : An adaptive optical radial basis function neural network classifier is experimentally demonstrated. We describe a spatially multiplexed system incorporating on-line adaptation of weights and basis function widths to provide robustness to optical system imperfections and system noise. The optical system computes the Euclidean distances between a 100-dimensional input vector and 198 stored reference patterns in parallel using dual vector-matrix multipliers and a contrast-reversing spatial light modulator. Software is used to emulate an analog electronic chip that performs the on-line learning of the weights and basis function widths. An experimental recognition rate of 92.7% correct out of 300 testing samples is achieved with the adaptive training versus 31.0% correct for non-adaptive training. We compare the experimental results with a detailed computer model of the system in order to analyze the influence of various noise sources on the system performance. (KAR) P. 3

Descriptors :   *NEURAL NETS, *COMPUTER ARCHITECTURE, *PATTERN RECOGNITION, *MULTIPLEXING, COMPUTER PROGRAMS, COMPUTERIZED SIMULATION, SOURCES, OPTICAL EQUIPMENT, REAL TIME, RATES, CHIPS(ELECTRONICS), PARALLEL PROCESSING, PROBLEM SOLVING, WEIGHT, NOISE, ONLINE SYSTEMS, ADAPTATION, LEARNING, ANALOG SYSTEMS, ADAPTIVE TRAINING.

Subject Categories : Computer Hardware
      Cybernetics

Distribution Statement : APPROVED FOR PUBLIC RELEASE