Single chip photonic deep neural network with accelerated training

TL;DR

This paper presents a scalable photonic integrated circuit that combines linear and nonlinear optical units for deep neural networks, enabling in situ training with high speed and energy efficiency.

Contribution

It introduces a fully integrated coherent optical neural network with in situ training, combining novel optical units for scalable, fast, and energy-efficient neural network processing.

Findings

Achieved 92.7% accuracy on vowel classification with in situ training.

Demonstrated optical neural network with 3 layers, 12 NOFUs, and 3 CMXUs.

Enabled inference at nanosecond latency and femtojoule energy per operation.

Abstract

As deep neural networks (DNNs) revolutionize machine learning, energy consumption and throughput are emerging as fundamental limitations of CMOS electronics. This has motivated a search for new hardware architectures optimized for artificial intelligence, such as electronic systolic arrays, memristor crossbar arrays, and optical accelerators. Optical systems can perform linear matrix operations at exceptionally high rate and efficiency, motivating recent demonstrations of low latency linear algebra and optical energy consumption below a photon per multiply-accumulate operation. However, demonstrating systems that co-integrate both linear and nonlinear processing units in a single chip remains a central challenge. Here we introduce such a system in a scalable photonic integrated circuit (PIC), enabled by several key advances: (i) high-bandwidth and low-power programmable nonlinear optical function units (NOFUs); (ii) coherent matrix multiplication units (CMXUs); and (iii) in situ training with optical acceleration. We experimentally demonstrate this fully-integrated coherent optical neural network (FICONN) architecture for a 3-layer DNN comprising 12 NOFUs and three CMXUs operating in the telecom C-band. Using in situ training on a vowel classification task, the FICONN achieves 92.7% accuracy on a test set, which is identical to the accuracy obtained on a digital computer with the same number of weights. This work lends experimental evidence to theoretical proposals for in situ training, unlocking orders of magnitude improvements in the throughput of training data. Moreover, the FICONN opens the path to inference at nanosecond latency and femtojoule per operation energy efficiency.