Abstract:The computational demands of modern AI have spurred interest in optical neural networks (ONNs) which offer the potential benefits of increased speed and lower power consumption. However, current ONNs face various challenges,most significantly a limited calculation precision (typically around 4 bits) and the requirement for high-resolution signal format converters (digital-to-analogue conversions (DACs) and analogue-to-digital conversions (ADCs)). These challenges are inherent to their analog computing nature and pose significant obstacles in practical implementation. Here, we propose a digital-analog hybrid optical computing architecture for ONNs, which utilizes digital optical inputs in the form of binary words. By introducing the logic levels and decisions based on thresholding, the calculation precision can be significantly enhanced. The DACs for input data can be removed and the resolution of the ADCs can be greatly reduced. This can increase the operating speed at a high calculation precision and facilitate the compatibility with microelectronics. To validate our approach, we have fabricated a proof-of-concept photonic chip and built up a hybrid optical processor (HOP) system for neural network applications. We have demonstrated an unprecedented 16-bit calculation precision for high-definition image processing, with a pixel error rate (PER) as low as $1.8\times10^{-3}$ at an signal-to-noise ratio (SNR) of 18.2 dB. We have also implemented a convolutional neural network for handwritten digit recognition that shows the same accuracy as the one achieved by a desktop computer. The concept of the digital-analog hybrid optical computing architecture offers a methodology that could potentially be applied to various ONN implementations and may intrigue new research into efficient and accurate domain-specific optical computing architectures for neural networks.

Numerical simulations on optoelectronic deep neural network hardware based on self-referential holography

Optical Analog Computing Devices Designed by Deep Neural Network

All-optical Diffractive Neural Networked Terahertz Hologram.

On-chip 4F-System Based on Concave Mirrors for Optical Neural Networks

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

1-bit Quantized On-chip Hybrid Diffraction Neural Network Enabled by Authentic All-optical Fully-connected Architecture

Digital-analog hybrid matrix multiplication processor for optical neural networks

Optoelectronic implementations of neural networks

Multilayer Optoelectronic Hybrid Convolutional Neural Network with an Optical 4F-System Recurrent Structure

Optical-electronic neural network for multi-modality and high-accurate lensless optics design and image reconstruction

Space-efficient optical computing with an integrated chip diffractive neural network

Deep learning with coherent VCSEL neural networks

Simulating an Integrated Photonic Image Classifier for Diffractive Neural Networks

Non-volatile Reconfigurable Digital Optical Diffractive Neural Network Based on Phase Change Material

Single-shot optical neural network

Sophisticated Deep Learning with On-Chip Optical Diffractive Tensor Processing

High-speed serial deep learning through temporal optical neurons

Optical Neural Networks for Holographic Image Recognition (invited Paper)

Multimodal deep learning using on-chip diffractive optics with in situ training capability

Optical Rectifying Linear Units for Back-Propagation Learning in a Deep Holographic Convolutional Neural Network

Optoelectronic recurrent neural network using optical-electrical-optical converters with RC delay