A spiking photonic neural network of 40 000 neurons, trained with latency and rank-order coding for leveraging sparsity

Tools

Talukder, Ria and Skalli, Anas and Porte, Xavier and Thorpe, Simon and Brunner, Daniel (2025) A spiking photonic neural network of 40 000 neurons, trained with latency and rank-order coding for leveraging sparsity. Neuromorphic Computing and Engineering, 5 (3). 034003. ISSN 2634-4386 (https://doi.org/10.1088/2634-4386/addee7)

[thumbnail of Talukder_2025_Neuromorph._Comput._Eng._5_034003]
Preview
 Text. Filename: Talukder_2025_Neuromorph._Comput._Eng._5_034003.pdf
Final Published Version
License: Creative Commons Attribution 4.0 logo
Download (1MB)| Preview

Abstract

Spiking neural networks (SNNs) are neuromorphic systems that emulate certain aspects of biological neural tissue, offering potential advantages in energy efficiency and speed by for example leveraging sparsity. While CMOS-based electronic SNN hardware has shown promise, scalability and parallelism challenges remain. Photonics provides a promising platform for SNNs due to the speed of excitable photonic devices standing in as neurons and the parallelism and low-latency of optical signal conduction. Here, we present a photonic SNN comprising 40 000 neurons using off-the-shelf components, including a spatial light modulator and a CMOS camera, enabling scalable and cost-effective implementations for photonic SNN proof of concept studies. The system is governed by a modified Ikeda map, where adding slow inhibitory feedback forcing introduces excitability akin to biological dynamics. Using latency encoding and sparsity, the network achieves 83.5% accuracy on MNIST handwritten digits using only 22% of neurons, and 77.5% with only 8.5% of neurons. Training is performed via liquid state machine concepts combined with the hardware-compatible simultaneous perturbation stochastic approximation algorithm, marking its first use in photonic neural networks. This demonstration integrates photonic nonlinearity, excitability, and sparse computation, paving the way for efficient large-scale photonic neuromorphic systems.

ORCID iDs

Talukder, Ria, Skalli, Anas, Porte, Xavier ORCID logoORCID: https://orcid.org/0000-0002-9869-7170, Thorpe, Simon and Brunner, Daniel;
  • Item type:Article
    ID code:93413
    Dates:
    Date
    Event
    1 September 2025
    Published
    4 July 2025
    Published Online
    30 May 2025
    Accepted
    28 November 2024
    Submitted
    Subjects:Science > Physics
    Department:Faculty of Science > Physics
    Depositing user: Pure Administrator
    Date deposited:07 Jul 2025 09:03
    Last modified:14 Aug 2026 10:37
    URI:https://strathprints.strath.ac.uk/id/eprint/93413
CORE (COnnecting REpositories)