A Robust MZI-Based Optical Neural Network Using QR Decomposition
Journal of Lightwave Technology · 01.02.2025 · 43 (3) · 1024-1031
DOI: 10.1109/jlt.2024.3476113 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Lightwave Technology, erschienen am 01.02.2025. Verfasst von Jian Lin und 9 weiteren Autor:innen. Die Arbeit wurde bisher 12-mal zitiert (Crossref) und nennt 26 Literaturangaben. Gefördert von: National Natural Science Foundation of China, Zhejiang Provincial Natural Science Foundation of China, Youth Science and Technology Innovation Leading Talent Project of Ningbo.
Abstract (Originalsprache)
In this paper, a robust MZI-based optical neural network using QR decomposition is proposed and investigated. To construct optical linear unit, (N2+N)/2 MZIs are required to achieve N × N weight matrix W in the case of QR decomposition, while in the case of singular value decomposition, N2 MZIs are needed. A two-layer MZI-based optical neural network using QR decomposition, in which each layer comprises the 4 × 4 optical linear unit and absolute activation function, is designed to identify the first four MNIST handwritten digit images to verify the feasibility of our proposal. A validation accuracy of 81.5% is obtained in the simulation. As a proof of concept, the designed MZI-based optical neural network using QR decomposition was fabricated an SOI platform. Experimental results show that, the measured validation accuracy is 72.5%. Under the same situation, the corresponding validation accuracy of the MZI-based optical neural network using singular value decomposition is estimated to be 67.25%. Compared to MZI-based optical neural network using singular value decomposition, the presented MZI-based optical neural network using QR decomposition needs fewer MZIs and has stronger robustness. If the size of the matrix W increases, these advantages of the presented MZI-based optical neural network using QR decomposition will become more apparent.
Förderung
- National Natural Science Foundation of China
- Zhejiang Provincial Natural Science Foundation of China
- Youth Science and Technology Innovation Leading Talent Project of Ningbo
- Natural Science Foundation of Ningbo
- K. C. Wong Magna Fund in Ningbo University
Zitieren
APA
Lin, J., Yang, K., Fu, Q., Wang, P., Dai, S., Chen, W., Kong, D., Li, J., Dai, T., & Yang, J. (2025). A Robust MZI-Based Optical Neural Network Using QR Decomposition. Journal of Lightwave Technology, 43(3), 1024-1031. https://doi.org/10.1109/jlt.2024.3476113
Vancouver
Lin J, Yang K, Fu Q, Wang P, Dai S, Chen W, et al. A Robust MZI-Based Optical Neural Network Using QR Decomposition. Journal of Lightwave Technology. 2025;43(3):1024-1031. doi:10.1109/jlt.2024.3476113
BibTeX
@article{Lin2025,
author = {Lin, Jian and Yang, Kang and Fu, Qiang and Wang, Pengjun and Dai, Shixun and Chen, Weiwei and Kong, Dejun and Li, Jun and Dai, Tingge and Yang, Jianyi},
title = {A Robust MZI-Based Optical Neural Network Using QR Decomposition},
journal = {Journal of Lightwave Technology},
year = {2025},
volume = {43},
number = {3},
pages = {1024--1031},
doi = {10.1109/jlt.2024.3476113}
}
EUVIMED