English

TRON: Trainable, architecture-reconfigurable random optical neural networks

Optics 2026-04-23 v2 Applied Physics

Abstract

Deep learning has triggered explosive growth in the demand for specialized hardware processors, thus motivating the development of scalable and reconfigurable computing substrates. Optical processors offer a fundamentally different computing paradigm, combining massive parallelism and ultrahigh bandwidth with the potential for substantial energy savings. However, progress has been constrained by the absence of scalable and reconfigurable architectures that can implement a broad class of network architectures. Here, we introduce TRON, a scalable and trainable optoelectronic deep optical neural network that exploits a multi-scattering medium and a DMD as a learnable, high-dimensional dense optical matrix multiplier, processing with fixed and tunable optical operations. We perform in-situ optimization of the optical parameters involved in the scattering process, together with automated neural architecture search (NAS) and optimization directly on optics. The experimental results demonstrate that in-situ NAS is essential to discover architectures that adapt to both the task and hardware constraints, establishing a viable path towards large-scale optical processors for next-generation machine learning and data-intensive computing.

Keywords

Cite

@article{arxiv.2604.16228,
  title  = {TRON: Trainable, architecture-reconfigurable random optical neural networks},
  author = {Ziao Wang and Fei Xia and Logan G. Wright and Tatsuhiro Onodera and Martin Stein and Jianqi Hu and Peter L. McMahon and Sylvain Gigan},
  journal= {arXiv preprint arXiv:2604.16228},
  year   = {2026}
}

Comments

14pages, 5 figures, 1 table

R2 v1 2026-07-01T12:14:40.032Z