All-optical doubly resonant cavities for energy-efficient ReLU function in nanophotonic deep learning
PLOS One · 17.06.2026 · 21 (6) · e0345850
DOI: 10.1371/journal.pone.0345850 ↗Kurz gefasst
Fachartikel in der Zeitschrift PLOS One, online erschienen am 17.06.2026. Verfasst von Amirreza Ahmadnejad und 2 weiteren Autor:innen. Beteiligte Einrichtung: Department of Electrical Engineering, Sharif University of Technology, Tehran, Iran. Hauptthemen laut MEDLINE: Deep Learning, Optics and Photonics. Die Arbeit nennt 91 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
This paper presents a novel approach to implementing all-optical Rectified Linear Unit (ReLU) activation functions using compact doubly-resonant cavities with dimensions of approximately 10 μ m. The proposed design leverages χ ( 2 ) nonlinear processes within carefully engineered photonic structures that simultaneously resonate at both fundamental and second-harmonic frequencies. By exploiting the phase-sensitive nature of second-harmonic generation, we demonstrate an optical analog to the ReLU function, achieving femtojoule-level activation energy—comparable to state-of-the-art approaches—while reducing device footprint by two orders of magnitude compared to previous implementations. The theoretical framework is developed using coupled-mode theory and validated through rigorous finite-difference time-domain simulations. Beyond ReLU, we show that the same physical structure can implement alternative activation functions such as ELU and GELU through simple adjustments to input conditions. Neural network simulations demonstrate that the proposed optical activation functions achieve classification accuracy within 0.4% of ideal electronic implementations while offering significant advantages in energy efficiency and processing speed. This work represents a significant advancement toward realizing energy-efficient, high-density optical neural networks for next-generation artificial intelligence hardware.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Autor:innen und Institutionen
- Department of Electrical Engineering, Sharif University of Technology, Tehran, Iran.
- Department of Computer Engineering, Sharif University of Technology, Tehran, Iran.
ORCID-Profile: Somayyeh Koohi
Zitieren
APA
Ahmadnejad, A., Asadi, M. M., & Koohi, S. (2026). All-optical doubly resonant cavities for energy-efficient ReLU function in nanophotonic deep learning. PLOS One, 21(6), e0345850. https://doi.org/10.1371/journal.pone.0345850
Vancouver
Ahmadnejad A, Asadi MM, Koohi S. All-optical doubly resonant cavities for energy-efficient ReLU function in nanophotonic deep learning. PLOS One. 2026;21(6):e0345850. doi:10.1371/journal.pone.0345850
BibTeX
@article{Ahmadnejad2026,
author = {Ahmadnejad, Amirreza and Asadi, Mohammad Mehrdad and Koohi, Somayyeh},
title = {All-optical doubly resonant cavities for energy-efficient ReLU function in nanophotonic deep learning},
journal = {PLOS One},
year = {2026},
volume = {21},
number = {6},
pages = {e0345850},
doi = {10.1371/journal.pone.0345850}
}
EUVIMED