Neural network implementation using self-lensing media

Abstract

An all-optical implementation of a feed-forward artificial neural network is presented that uses self-lensing materials in which the index of refraction is irradiance dependent. Many of these types of material have ultrafast response times and permit both weighted connections and nonlinear neuron processing to be implemented with only thin material layers separated by free space. Both neuron processing and weighted interconnections emerge directly from the physical optics of the device. One creates virtual neurons and their connections simply by applying patterns of irradiance to thin layers of the nonlinear media. This is a result of a variation of the refractive-index profile of the self-lensing nonlinear media in response to the applied irradiance. An optical-backpropagation training method for this network is presented. The optical backpropagation is a training method that can be implemented potentially within the same optical device as the forward calculations, although several issues crucial to this po sibility remain to be addressed. Such a network was numerically simulated and trained to solve many benchmark classification problems, and some of these results are presented. To demonstrate the feasibility of building such a network, we also describe experimental work in the construction of an optical network trained to perform a logic XNOR function. This network, as a proof of concept, uses a relatively slow thermal nonlinear material with ~1-s response time.

Reinforcement and backpropagation training for an optical neural network using self-lensing effects

“…The thermal material gives a slow response to changes in the irradiance profile of the input and weighting beam. Yet, the material's simplicity of handling and low power requirement allow for the implementation and testing in real time of the networks described in Skinner et al [4], [5]. The resolution of the applied weights in the material is limited by its thermal diffusivity because heat transfer will occur between adjacent bright and dark regions reducing the effective optical resolution of the material.…”

“…Backpropagation, on the other hand, requires no increase in training epochs as the number of layers increases. To implement error backpropagation for this type of ANN, a set of equations [4], [5] was developed. The optical backpropagation algorithm presented by Steck et al [4], and Skinner et al [5] required further development in order to be implemented in optical hardware.…”

“…Skinner et al [4], [5] proposed a network that uses self-lensing nonlinear optical materials to steer light with light by using the optical concepts of cross-phase and self-phase modulation [6]. Computer simulations have demonstrated that this network can be trained to solve a variety of benchmark problems including data classification, sonar echo classification and robot control [4], [5].…”

“…Computer simulations have demonstrated that this network can be trained to solve a variety of benchmark problems including data classification, sonar echo classification and robot control [4], [5]. In this paper, an experimental realization of this feedforward neural network was trained using reinforcement and backpropagation algorithms.…”

“…1. It is a sandwich of two thick linear layers (free space) and a thin layer of a self-defocusing thermal nonlinear material (NLM) [4], [5]. A laser beam, called the forward information beam, interconnects all layers.…”