Phase-limited quantization-aware training for diffractive deep neural networks
Applied Optics · 20.02.2025 · 64 (6) · 1413
DOI: 10.1364/ao.548035 ↗Kurz gefasst
Fachartikel in der Zeitschrift Applied Optics, online erschienen am 11.02.2025. Verfasst von Yu Wang und 2 weiteren Autor:innen. Die Arbeit wurde bisher 4-mal zitiert (Crossref) und nennt 22 Literaturangaben. Gefördert von: Xiong'an New Area Science and Technology Innovation Project, National Key Research and Development Program of China.
Abstract (Originalsprache)
In recent years, all-optical diffractive deep neural networks (D 2 NNs) have demonstrated exceptional performances in many fields. Discretizing the grating height can reduce the complexity and enhance the network manufacturing efficiency. For this purpose, we propose a phase-limited quantization-aware training (PLQAT) method and construct an all-optical D 2 NN to discretize the network for the MNIST image classification task. Our results indicate that the PLQAT method improves the classification performance of the D 2 NN by 0.11–27.96% across different bit levels compared to the classical algorithm. Furthermore, we identified 3-bit quantization as the optimal choice, discretized the phase values in five layers to eight levels within [0,2 π ], and achieved a test accuracy of 96.22%. This method discretizes the height of gratings of D 2 NNs, effectively reducing the difficulty of grating etching while maintaining good network performance.
Förderung
- Xiong'an New Area Science and Technology Innovation Project
- National Key Research and Development Program of China
Zitieren
APA
Wang, Y., Sha, Q., & Qi, F. (2025). Phase-limited quantization-aware training for diffractive deep neural networks. Applied Optics, 64(6), 1413. https://doi.org/10.1364/ao.548035
Vancouver
Wang Y, Sha Q, Qi F. Phase-limited quantization-aware training for diffractive deep neural networks. Applied Optics. 2025;64(6):1413. doi:10.1364/ao.548035
BibTeX
@article{Wang2025,
author = {Wang, Yu and Sha, Qi and Qi, Feng},
title = {Phase-limited quantization-aware training for diffractive deep neural networks},
journal = {Applied Optics},
year = {2025},
volume = {64},
number = {6},
pages = {1413},
doi = {10.1364/ao.548035}
}
EUVIMED