Inverse design of convolutional neural networks via nanophotonic kernels
Kaida Tang and
Ran Hao
Journal of Electromagnetic Waves and Applications, 2024, vol. 38, issue 10, 1176-1186
Abstract:
We proposed an ultra-compact neural network based on inversely designed photonic kernels, which has the advantages of a small footprint, high computing speed, and low power consumption. The core layers of the neural network, composed of automatically optimized photonic multiplication kernels and addition kernels, can perform arbitrary photonic computations at the speed of the light wave, offering more advantages than their electronic counterparts. Furthermore, the convolution, pooling, and fully-connected layers can be realized by the appropriate permutation of the proposed optical kernels. We constructed an all-optical convolutional neural network with a high accuracy of 97.7% for recognizing handwritten digits. Our results could significantly promote the inverse design of photonic neural networks.
Date: 2024
References: Add references at CitEc
Citations:
Downloads: (external link)
http://hdl.handle.net/10.1080/09205071.2024.2360554 (text/html)
Access to full text is restricted to subscribers.
Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.
Export reference: BibTeX
RIS (EndNote, ProCite, RefMan)
HTML/Text
Persistent link: https://EconPapers.repec.org/RePEc:taf:tewaxx:v:38:y:2024:i:10:p:1176-1186
Ordering information: This journal article can be ordered from
http://www.tandfonline.com/pricing/journal/tewa20
DOI: 10.1080/09205071.2024.2360554
Access Statistics for this article
Journal of Electromagnetic Waves and Applications is currently edited by Mohamad Abou El-Nasr and Pankaj Kumar Choudhury
More articles in Journal of Electromagnetic Waves and Applications from Taylor & Francis Journals
Bibliographic data for series maintained by Chris Longhurst ().