High Clockrate Free-space Optical In-Memory Computing

2025/09/23 by Liang, Yuanhao, Wang, James, Xue, Kaiwen +12 · 1 citation
#Emerging Technologies (cs.ET) #FOS: Computer and information sciences #FOS: Physical sciences #Optics (physics.optics)

paper · doi:10.48550/arxiv.2509.19642

Abstract

The ability to process and act on data in real time is increasingly critical for applications ranging from autonomous vehicles, three-dimensional environmental sensing and remote robotics. However, the deployment of deep neural networks (DNNs) in edge devices is hindered by the lack of energy-efficient scalable computing hardware. Here, we introduce a fanout spatial time-of-flight optical neural network (FAST-ONN) that calculates billions of convolutions per second with ultralow latency and power consumption. This is enabled by the combination of high-speed dense arrays of vertical-cavity surface-emitting lasers (VCSELs) for input modulation with spatial light modulators of high pixel counts for in-memory weighting. In a three-dimensional optical system, parallel differential readout allows signed weight values accurate inference in a single shot. The performance is benchmarked with feature extraction in You-Only-Look-Once (YOLO) for convolution at 100 million frames per second (MFPS), and in-system backward propagation training with photonic reprogrammability. The VCSEL transmitters are implementable in any free-space optical computing systems to improve the clockrate to over gigahertz. The high scalability in device counts and channel parallelism enables a new avenue to scale up free space computing hardware.

Citations

• Edge Computing and its Application in Robotics: A Survey
• Integrated lithium niobate photonic computing circuit based on efficient and high-speed electro-optic conversion
• Optical Generative Models
• 120 GOPS Photonic Tensor Core in Thin-film Lithium Niobate for Inference and in-situ Training
• Roadmapping the Next Generation of Silicon Photonics
• TPU v4: An Optically Reconfigurable Supercomputer for Machine Learning with Hardware Support for Embeddings
• Single-Shot Optical Neural Network
• Metasurface-Enabled On-Chip Multiplexed Diffractive Neural Networks in the Visible
• Impact of Thermal Throttling on Long-Term Visual Inference in a\n CPU-based Edge Device
• Massively Parallel Amplitude-Only Fourier Neural Network
• Hexagonal Transverse Coupled Cavity VCSEL Redefining the High-Speed Lasers
• Accelerating convolutional neural network by exploiting sparsity on GPUs
• Optimization problems with low SWaP tactical Computing
• All-optical machine learning using diffractive deep neural networks
• Efficient Processing of Deep Neural Networks: A Tutorial and Survey
• Efficient Processing of Deep Neural Networks: A Tutorial and Survey
• Neuromorphic photonic networks using silicon photonic weight banks
• Deep learning with coherent nanophotonic circuits
• Quantized Neural Networks: Training Neural Networks with Low Precision Weights and Activations
• Deep Residual Learning for Image Recognition
• You Only Look Once: Unified, Real-Time Object Detection
• U-Net: Convolutional Networks for Biomedical Image Segmentation
• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Microsoft COCO: Common Objects in Context
• A flexible new technique for camera calibration

Cited by

• Homodyne Photonic Tensor Processor exceeds 1,000-TOPS

Related

• Single-Shot Optical Neural Network
• Model-free front-to-end training of a large high performance laser neural network
• High-throughput optical neural networks based on temporal computing
• A complete, parallel and autonomous photonic neural network in a semiconductor multimode laser
• Massively Parallel Amplitude-Only Fourier Neural Network