GaAs/AlGaAs optical synaptic interconnection device for neural networks

Abstract

A GaAs/AlGaAs optical synaptic interconnection device for neural networks is reported for the first time to our knowledge. This device consists of a light-emitting-diode array, an interconnection matrix, and a photodiode array, which are integrated into a hybrid-layered structure on a GaAs substrate. The device structure and characteristics are reported in detail. The fabricated device can simulate a 32-neuron system. Experimental results of the Hopfield associative memory with three stored vectors are also described.

Learning Pattern Recognition Through Quasi-Synchronization of Phase Oscillators

“…In Hopfield's model, learning happens by changes in the connections between each node. Because it is fully connected, it is very hard to produce computer chips that mimic large-scale Hopfield networks [42]. It is possible that, by having fixed connections but changing natural frequencies, our model may face less difficulties with hardware implementation.…”

Optical implementation of hopfield associative memory based on imaging system using a liquid crystal device

“…The operation principle of this system is simple and, in principle. quite similar to the sandwich structure proposed by Ohta et al [13]. But this system has some new features: The light intensity energy detected by CCD is enhanced by use of lens 2; the input pattern size on LCD 1, the interconnection matrix size on LCD 2 .…”

“…How can this 4D interconnection matrix be arranged in 2D plane? Jang, Shin, and Lee [15] have shown two equivalent methods; one is the arrangement order widely used by the lenslet method [13]; the other is shown in Figure 2 , and was used in this experiment. When the characters A , B, L (7 X 7 pixels) are stored, the positive and negative parts of the 4D interconnection matrix, arranged as in the manner of Figure 2, are respectively displayed on LCD 2 one after the other.…”

<title>Optical learning neural networks with two-dimensional structure</title>

“…(1) can be rewritten as Xkl (jld + a ) x I} + (1 -O=f(X-Na), (4) where N is the number of input neurons. (1) can be rewritten as Xkl (jld + a ) x I} + (1 -O=f(X-Na), (4) where N is the number of input neurons.…”