Optical Realizations of Hopfield and Boltzmann Neural Networks
- 170 Downloads
One of the major reasons why Hopfield neural networks attract much interest as candidates for optical realizations is the fact that they easily provide schemes for pattern recognition and associative memories. Beneath the possibility to exploit the associative memory properties of classical Van der Lugt filters as they have been discussed in chapter 7, models based on directly transferring the Hopfield model from their digital electronic manifestations into optical systems are the most widely spread and successfully realized ones.
KeywordsSimulated Annealing Optical Realization Associative Memory Synaptic Weight Hopfield Neural Network
Unable to display preview. Download preview PDF.
- 1.B.K. Jenkins, A.R. Tanguay, Photonic Implementations of Neural Networks, Chapter 9 in Neural Networks for Signal Processing, B. Kosko, Ed., Prentice-Hall (1992).Google Scholar
- 2.N.H. Farhat, D. Psaltis, Optical Impelemtation of Associative Memory Based on Models of Neural Networks, in Optical Signal Processing, Ed. J.L. Horner, Academic Press (1987).Google Scholar