Comparison of Models for Training Optical Matrix Multipliers in Neuromorphic PICs

TL;DR

This paper compares physics-based and neural-network-based models for training optical matrix multipliers in neuromorphic photonic integrated circuits, demonstrating the neural network's superior performance in handling thermal crosstalk.

Contribution

It introduces a comparative analysis showing neural-network models outperform physics-based models for photonic chip training, especially under thermal crosstalk conditions.

Findings

Neural-network models achieve higher testing accuracy.

Neural models better handle thermal crosstalk effects.

Physics-based models are less accurate in complex conditions.

Abstract

We experimentally compare simple physics-based vs. data-driven neural-network-based models for offline training of programmable photonic chips using Mach-Zehnder interferometer meshes. The neural-network model outperforms physics-based models for a chip with thermal crosstalk, yielding increased testing accuracy.