Photonic convolutional neural networks using integrated diffractive optics

Page view(s)
6
Checked on Sep 26, 2022
Photonic convolutional neural networks using integrated diffractive optics
Title:
Photonic convolutional neural networks using integrated diffractive optics
Other Titles:
IEEE Journal of Selected Topics in Quantum Electronics
Publication Date:
27 March 2020
Citation:
J. R. Ong, C. C. Ooi, T. Y. L. Ang, S. T. Lim and C. E. Png, "Photonic convolutional neural networks using integrated diffractive optics," in IEEE Journal of Selected Topics in Quantum Electronics.
Abstract:
With recent rapid advances in photonic integrated circuits, it has been demonstrated that programmable photonic chips can be used to implement artificial neural networks. Convolutional neural networks (CNN) are a class of deep learning methods that have been highly successful in applications such as image classification and speech processing. We present an architecture to implement a photonic CNN using the Fourier transform property of integrated star couplers. We show, in computer simulation, high accuracy image classification using the MNIST dataset. We also model component imperfections in photonic CNN and show that the performance degradation can be recovered in a programmable chip. Our proposed architecture provides a large reduction in physical footprint compared to current implementations as it utilizes the natural advantages of optics and hence offers a scalable pathway towards integrated photonic deep learning processors.
License type:
Publisher Copyright
Funding Info:
This research is supported by core funding from: Institute of High Performance Computing
Grant Reference no. : SC23-R0005
Description:
© 2020 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.
ISSN:
1077-260X
1558-4542
Files uploaded:

File Size Format Action
arxiv-maintext.pdf 533.09 KB PDF Open