Programmable diffractive neural networks enabled by microfluidic refractive-index encoding
Optics Letters · 15.11.2026 · 51 (22) · 6207
DOI: 10.1364/ol.614327 ↗Kurz gefasst
Leserbrief in der Zeitschrift Optics Letters, online erschienen am 03.11.2026. Verfasst von Shengting Zhu und 1 weiteren Autor:innen. Beteiligte Einrichtung: School of Artificial Intelligence Science and Technology, University of Shanghai for Science and Technology, Shanghai 200093, China. Die Arbeit nennt 37 Literaturangaben. Gefördert von: Science and Technology Commission of Shanghai Municipality, Shanghai Frontiers Science Center Program, Program for Professor of Special Appointment (Eastern Scholar) at Shanghai Institutions of Higher Learning.
Abstract (Originalsprache)
Diffractive deep neural networks (D 2 NNs) provide a compact platform for optical computing, but their functions are generally fixed after fabrication. Here, we introduce fluidic refractive-index control as an additional degree of freedom for programmable D 2 NNs. By integrating microfluidic chambers with a proposed monolithic D 2 NN architecture, the optical response can be dynamically tuned without altering the physical structure. We demonstrate refractive-index-controlled holographic multiplexing and reconfigurable D 2 NN functions, including imaging, edge detection, and classification, through numerical simulations. Independent control of individual layers further enables binary encoding, allowing a three-layer network to realize eight computational states with high recognition accuracy. This approach provides a compact and scalable route toward reconfigurable optical computing and photonic intelligence.
Autor:innen und Institutionen
- School of Artificial Intelligence Science and Technology, University of Shanghai for Science and Technology, Shanghai 200093, China.
- Institute of Photonic Chips, University of Shanghai for Science and Technology, Shanghai 200093, China.
- Shanghai Key Laboratory of Neuromorphic Optical Computing and Edge Intelligence, University of Shanghai for Science and Technology, Shanghai 200093, China.
ORCID-Profile: Yinan Zhang
Förderung
- Science and Technology Commission of Shanghai Municipality
- Shanghai Frontiers Science Center Program
- Program for Professor of Special Appointment (Eastern Scholar) at Shanghai Institutions of Higher Learning
- National Natural Science Foundation of China
- Shanghai Science and Technology Program
Zitieren
APA
Zhu, S., & Zhang, Y. (2026). Programmable diffractive neural networks enabled by microfluidic refractive-index encoding. Optics Letters, 51(22), 6207. https://doi.org/10.1364/ol.614327
Vancouver
Zhu S, Zhang Y. Programmable diffractive neural networks enabled by microfluidic refractive-index encoding. Optics Letters. 2026;51(22):6207. doi:10.1364/ol.614327
BibTeX
@article{Zhu2026,
author = {Zhu, Shengting and Zhang, Yinan},
title = {Programmable diffractive neural networks enabled by microfluidic refractive-index encoding},
journal = {Optics Letters},
year = {2026},
volume = {51},
number = {22},
pages = {6207},
doi = {10.1364/ol.614327}
}
EUVIMED