Abstract
We demonstrate a universal optical vector convolutional accelerator operating at 11 Tera-OPS, generating convolutions of images of 250,000 pixels with 8-bit resolution for 10 kernels simultaneously. We use the same hardware to form a deep optical CNN with ten output neurons, achieving successful recognition of full 10 digits with 88% accuracy. Our approach is scalable and trainable for applications to unmanned vehicle and real-time video recognition.
© 2022 The Author(s)
PDF Article | Presentation VideoMore Like This
M. Tan, X. Xu, J. Wu, A. Boes, B. Corcoran, T. Nguyen, S. T. Chu, B. E. Little, R. Morandotti, A.J. Lowery, A. Mitchell, D. Hicks, and D. J. Moss
SpM5C.1 Signal Processing in Photonic Communications (SPPCom) 2021
M. Tan, X. Xu, J. Wu, A. Boes, B. Corcoran, T. Nguyen, S. T. Chu, B. E. Little, R. Morandotti, D. G. Hicks, Mitchell, and David J. Moss
Tu5B.1 Photonics in Switching and Computing (PS) 2021
Yang Sun, Mengxi Tan, Xingyuan Xu, Jiayang Wu, Yang Li, Sai T. Chu, Brent E. Little, Roberto Morandotti, Arnan Mitchell, and David J. Moss
SM1P.1 CLEO: Science and Innovations (CLEO:S&I) 2023