QuATON: Quantization Aware Training of Optical Neurons

TL;DR

This paper introduces a physics-informed quantization-aware training framework for optical neurons, enabling the design of robust optical processors with limited precision suitable for 3D fabrication, advancing optical computing capabilities.

Contribution

It presents the first training method that incorporates physical constraints and quantization levels for optical neurons, improving the robustness and feasibility of optical processors.

Findings

Designs state-of-the-art optical processors with quantized parameters

Demonstrates robustness of optical neural networks under physical constraints

Enables future 3D fabrication of optical processors with improved performance

Abstract

Optical processors, built with "optical neurons", can efficiently perform high-dimensional linear operations at the speed of light. Thus they are a promising avenue to accelerate large-scale linear computations. With the current advances in micro-fabrication, such optical processors can now be 3D fabricated, but with a limited precision. This limitation translates to quantization of learnable parameters in optical neurons, and should be handled during the design of the optical processor in order to avoid a model mismatch. Specifically, optical neurons should be trained or designed within the physical-constraints at a predefined quantized precision level. To address this critical issues we propose a physics-informed quantization-aware training framework. Our approach accounts for physical constraints during the training process, leading to robust designs. We demonstrate that our approach can design state of the art optical processors using diffractive networks for multiple physics based tasks despite quantized learnable parameters. We thus lay the foundation upon which improved optical processors may be 3D fabricated in the future.