OplixNet: Towards Area-Efficient Optical Split-Complex Networks with Real-to-Complex Data Assignment and Knowledge Distillation

TL;DR

OplixNet introduces a novel optical neural network framework that utilizes both amplitude and phase information for data processing, significantly reducing area requirements while maintaining acceptable accuracy levels.

Contribution

The paper proposes a new method to compress optical neural networks by leveraging real-to-complex data assignment and mutual learning, enhancing area efficiency.

Findings

75.03% area reduction with 0.33% accuracy loss on FCNN

74.88% area reduction with 2.38% accuracy loss on ResNet-32

Effective use of phase information improves area efficiency in ONNs

Abstract

Having the potential for high speed, high throughput, and low energy cost, optical neural networks (ONNs) have emerged as a promising candidate for accelerating deep learning tasks. In conventional ONNs, light amplitudes are modulated at the input and detected at the output. However, the light phases are still ignored in conventional structures, although they can also carry information for computing. To address this issue, in this paper, we propose a framework called OplixNet to compress the areas of ONNs by modulating input image data into the amplitudes and phase parts of light signals. The input and output parts of the ONNs are redesigned to make full use of both amplitude and phase information. Moreover, mutual learning across different ONN structures is introduced to maintain the accuracy. Experimental results demonstrate that the proposed framework significantly reduces the areas of ONNs with the accuracy within an acceptable range. For instance, 75.03% area is reduced with a 0.33% accuracy decrease on fully connected neural network (FCNN) and 74.88% area is reduced with a 2.38% accuracy decrease on ResNet-32.