Scalable Back-Propagation-Free Training of Optical Physics-Informed Neural Networks
ACM Transactions on Design Automation of Electronic Systems · 31.03.2027 · 32 (2) · 1-41
DOI: 10.1145/3837864 ↗Kurz gefasst
Fachartikel in der Zeitschrift ACM Transactions on Design Automation of Electronic Systems, online erschienen am 13.11.2026. Verfasst von Yequan Zhao und 8 weiteren Autor:innen. Die Arbeit nennt 94 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0).
Abstract (Originalsprache)
Physics intelligence and digital twins often require rapid and repeated performance evaluation of various engineering systems (e.g., robots, autonomous vehicles, semiconductor chips) to enable (almost) real-time actions or decision making. This has motivated the development of accelerated partial differential equation (PDE) solvers, in resource-constrained scenarios if the PDE solvers are to be deployed on the edge. Physics-informed neural networks (PINNs) have shown promise in solving high-dimensional PDEs, but the training time on state-of-the-art digital hardware (e.g., GPUs) is still orders-of-magnitude longer than the latency required for enabling real-time decision making. Photonic computing offers a potential solution to address this huge latency gap because of its ultra-high operation speed. However, the lack of photonic memory and the large device sizes prevent training real-size PINNs on photonic chips. This article proposes a completely back-propagation-free (BP-free) and highly scalable framework for training real-size PINNs on silicon photonic platforms. Our approach involves three key innovations: (1) a sparse-grid Stein derivative estimator to avoid the BP in the loss evaluation of a PINN, (2) a dimension-reduced zeroth-order optimization via tensor-train decomposition to achieve better scalability and convergence in BP-free training, and (3) a scalable on-chip photonic PINN training accelerator design using photonic tensor cores. We validate our numerical methods on both low- and high-dimensional PDE benchmarks. Through pre-silicon simulation based on real device parameters, we further demonstrate the significant performance benefit (e.g., real-time training, huge chip area reduction) of our photonic accelerator. Our code is available at https://github.com/olokevin/scalable_bpfree_onn_training .
Zitieren
APA
Zhao, Y., Yu, X., Xiao, X., Chen, Z., Liu, Z., Kurczveil, G., Beausoleil, R., Liu, S., & Zhang, Z. (2027). Scalable Back-Propagation-Free Training of Optical Physics-Informed Neural Networks. ACM Transactions on Design Automation of Electronic Systems, 32(2), 1-41. https://doi.org/10.1145/3837864
Vancouver
Zhao Y, Yu X, Xiao X, Chen Z, Liu Z, Kurczveil G, et al. Scalable Back-Propagation-Free Training of Optical Physics-Informed Neural Networks. ACM Transactions on Design Automation of Electronic Systems. 2027;32(2):1-41. doi:10.1145/3837864
BibTeX
@article{Zhao2027,
author = {Zhao, Yequan and Yu, Xinling and Xiao, Xian and Chen, Zhixiong and Liu, Ziyue and Kurczveil, Geza and Beausoleil, Raymond and Liu, Sijia and Zhang, Zheng},
title = {Scalable Back-Propagation-Free Training of Optical Physics-Informed Neural Networks},
journal = {ACM Transactions on Design Automation of Electronic Systems},
year = {2027},
volume = {32},
number = {2},
pages = {1--41},
doi = {10.1145/3837864}
}
EUVIMED