Pruning Coherent Integrated Photonic Neural Networks Using the Lottery Ticket Hypothesis

TL;DR

This paper introduces a hardware-aware pruning method for singular-value-decomposition-based coherent integrated photonic neural networks, significantly reducing their size and power consumption while maintaining accuracy.

Contribution

It applies the lottery ticket hypothesis to SC-IPNNs, enabling effective pruning and power savings not achievable with conventional methods.

Findings

Up to 89% of phase angles can be pruned with less than 5% accuracy loss.

Static power consumption reduced by up to 86%.

Pruning does not significantly degrade neural network performance.

Abstract

Singular-value-decomposition-based coherent integrated photonic neural networks (SC-IPNNs) have a large footprint, suffer from high static power consumption for training and inference, and cannot be pruned using conventional DNN pruning techniques. We leverage the lottery ticket hypothesis to propose the first hardware-aware pruning method for SC-IPNNs that alleviates these challenges by minimizing the number of weight parameters. We prune a multi-layer perceptron-based SC-IPNN and show that up to 89% of the phase angles, which correspond to weight parameters in SC-IPNNs, can be pruned with a negligible accuracy loss (smaller than 5%) while reducing the static power consumption by up to 86%.