High-order tensor flow processing using integrated photonic circuits
arXiv:2112.12322 · doi:10.1038/s41467-022-35723-2
Abstract
Tensor analytics lays mathematical basis for the prosperous promotion of multiway signal processing. To increase computing throughput, mainstream processors transform tensor convolutions to matrix multiplications to enhance parallelism of computing. However, such order-reducing transformation produces data duplicates and consumes additional memory. Here, we demonstrate an integrated photonic tensor flow processor without tensor-matrix transformation, which outputs the convolved tensor as the input tensor 'flows' through the processor. The hybrid manipulation of optical dimensions of wavelength, time, and space enables the direct representation and processing of high-order tensors in optical domain. In the proof-of-concept experiment, processing of multi-channel images and videos is accomplished at the frequency of 20 GHz. A convolutional neural network is demonstrated on the processor, which achieves an accuracy of 97.9 percent on action recognition.
References in corpus (9)
- Parallel convolution processing using an integrated photonic tensor core
- Photonics for artificial intelligence and neuromorphic computing
- 11 TeraFLOPs per second photonic convolutional accelerator for deep learning optical neural networks
- Single-chip photonic deep neural network for instantaneous image classification
- High-yield wafer-scale fabrication of ultralow-loss, dispersion-engineered silicon nitride photonic circuits
- Programmable Phase-change Metasurfaces on Waveguides for Multimode Photonic Convolutional Neural Network
- High-order tensor flow processing using integrated photonic circuits
- Silicon microring synapses enable photonic deep learning beyond 9-bit precision
- Optical coherent dot-product chip for sophisticated deep learning regression