Deep Learning with Coherent VCSEL Neural Networks

Abstract

Deep neural networks (DNNs) are reshaping the field of information processing. With their exponential growth challenging existing electronic hardware, optical neural networks (ONNs) are emerging to process DNN tasks in the optical domain with high clock rates, parallelism and low-loss data transmission. However, to explore the potential of ONNs, it is necessary to investigate the full-system performance incorporating the major DNN elements, including matrix algebra and nonlinear activation. Existing challenges to ONNs are high energy consumption due to low electro-optic (EO) conversion efficiency, low compute density due to large device footprint and channel crosstalk, and long latency due to the lack of inline nonlinearity. Here we experimentally demonstrate an ONN system that simultaneously overcomes all these challenges. We exploit neuron encoding with volume-manufactured micron-scale vertical-cavity surface-emitting laser (VCSEL) transmitter arrays that exhibit high EO conversion (<5 attojoule/symbol with $V_{\pi }$=4 mV), high operation bandwidth (up to 25 GS/s), and compact footprint (<0.01 mm${}^2$ per device). Photoelectric multiplication allows low-energy matrix operations at the shot-noise quantum limit. Homodyne detection-based nonlinearity enables nonlinear activation with instantaneous response. The full-system energy efficiency and compute density reach 7 femtojoules per operation (fJ/OP) and 25 TeraOP/(mm${}^2\cdot$ s), both representing a >100-fold improvement over state-of-the-art digital computers, with substantially several more orders of magnitude for future improvement. Beyond neural network inference, its feature of rapid weight updating is crucial for training deep learning models. Our technique opens an avenue to large-scale optoelectronic processors to accelerate machine learning tasks from data centers to decentralized edge devices.

10 pages, 5 figures

References in corpus (6)

• PyTorch: An Imperative Style, High-Performance Deep Learning Library
• 11 TeraFLOPs per second photonic convolutional accelerator for deep learning optical neural networks
• Machine learning for molecular simulation
• Attojoule Optoelectronics for Low-Energy Information Processing and Communications: a Tutorial Review
• Delocalized Photonic Deep Learning on the Internet's Edge
• All-optical ultrafast ReLU function for energy-efficient nanophotonic deep learning

Cited by in corpus (22)

• Photonics for Neuromorphic Computing: Fundamentals, Devices, and Opportunities
• Nonlinear optical encoding enabled by recurrent linear scattering
• Experimental reservoir computing with diffractively coupled VCSELs
• All-optical nonlinear activation function based on stimulated Brillouin scattering
• An optoacoustic field-programmable perceptron for recurrent neural networks
• Transferable polychromatic optical encoder for neural networks
• Waveguide-multiplexed photonic matrix-vector multiplication processor using multiport photodetectors
• Ultrafast one-chip optical receiver with functional metasurface
• Unitary control of partially coherent waves. II. Transmission or reflection
• Controllable distant interactions at bound state in the continuum
• Single-Shot Matrix-Matrix Multiplication Optical Tensor Processor for Deep Learning
• High-speed coherent photonic random-access memory in long-lasting sound waves
• A 262 TOPS Hyperdimensional Photonic AI Accelerator powered by a Si3N4 microcomb laser
• SUANPAN: Scalable Photonic Linear Vector Machine
• Model-free Optical Processors using In Situ Reinforcement Learning with Proximal Policy Optimization
• Photonic systolic array for all-optical matrix-matrix multiplication
• Enhanced Polarization Locking in VCSELs
• Compute-first optical detection for noise-resilient visual perception
• Quantum-secure multiparty deep learning
• Buried Stressor Engineering for Position-Controlled InGaAs Quantum Dots with Local Density Variation for Integrated Quantum Photonics
• Implementation of transformer-based LLMs with large-scale optoelectronic neurons on a CMOS compatible platform
• Ultrafast neural sampling with spiking nanolasers