Hybrid training of optical neural networks

Abstract

Optical neural networks are emerging as a promising type of machine learning hardware capable of energy-efficient, parallel computation. Today’s optical neural networks are mainly developed to perform optical inference after in silico training on digital simulators. However, various physical imperfections that cannot be accurately modeled may lead to the notorious “reality gap” between the digital simulator and the physical system. To address this challenge, we demonstrate hybrid training of optical neural networks where the weight matrix is trained with neuron activation functions computed optically via forward propagation through the network. We examine the efficacy of hybrid training with three different networks: an optical linear classifier, a hybrid opto-electronic network, and a complex-valued optical network. We perform a study comparative to in silico training, and our results show that hybrid training is robust against different kinds of static noise. Our platform-agnostic hybrid training scheme can be applied to a wide variety of optical neural networks, and this work paves the way towards advanced all-optical training in machine intelligence.

A Compact Butterfly-Style Silicon Photonic–Electronic Neural Chip for Hardware-Efficient Deep Learning

“…More details of our hardware-aware training framework are shown in the Supporting Information. Compared to prior on-chip training protocols based on derivative-free optimization algorithms ,,, and gradient approximation using ideal simulation models, − our AI-assisted ONN learning shows considerably higher scalability and effectiveness in robust optical neural chip training.…”

Free-space optical multiplexed orbital angular momentum beam identification system using Fourier optical convolutional layer based on 4f system

“…[52][53][54][55][56][57] Learning-based Imaging and focusing through scattering medium, computer-generated holography based on machine learning have been studied. [58][59][60][61][62][63][64][65][66][67][68][69][70][71][72][73][74][75] In FSOC, Timothy Doster and Abbie T. Watnik firstly introduced new method to detect the active OAM modes in a transmission link by utilizing convolutional neural network (CNN) [1]. Qinghua Tian et al proposed a novel turbo-coded 16-ary orbital angular momentum-shift keying -free space optical (OAM-SK-FSO) communication system combining a CNN based adaptive demodulator under strong AT for the first time.…”

Training on the optical system: local search method

“…[76][77][78][79][80][81][82] Due to the lack of monitoring for intermediate status, implementing gradient descent algorithms 83 for training optical neural networks on real optical systems can prove to be difficult. [84][85][86][87][88][89][90][91][92][93][94][95] This presents a challenge for achieving optimal performance in real-world scenarios. To overcome this issue, the local search algorithm [96][97][98][99][100] has been adapted for training optical neural networks on experimental setups.…”