Electronic circuits are not the only type of device that can exhibit behavior able to replicate biological neurons. In the past 20 years, researchers have developed optical neurons that also exhibit biomimetic properties and can therefor be considered in the context of photonic neuromorphic

computing. Such systems are much faster (< 1 nsec spikes) and very low energy. Their field of applications are therefore likely to be very different from organic neuromorphic hardware. Photonic neuromorphic circuits are still at a very early stage of development as the technology to integrate integrated arrays is still quite difficult.

The photonic circuits are based on microcavity lasers. The laser consists of an active zone of quantum wells  sandwiched between two Bragg reflectors. The active zone allows for stimulated emission (gain) when pumped with a laser of the appropriate frequency. In addition, it contains a saturable absorber, whose absorption saturates with the intracavity laser intensity. The ‘pump’ of the laser can be through of as an ion channel allowing an increase in gain, the  saturable absorption can be thought  of as allowing a decrease in gain. The result is a dyanmic system whose properties can ressemble a spiking biological neuron.

The role of LPICM in this project is to develop classification schemes by which the optical neurons could be used to solve targeted tasks. One project develops the methodology to realize an extreme learning machine type classification using photochromic molecules, in collaboration with chemists. Another project considers how the optical neurons can be used to realize time to first spike encoding architectures. The project is financed by two ANR grants: Photomic (2022-2026) and ANACONDA (2020-2023) and grant from LABEX Nanosaclay CANAPO (2021-2023).

See our collaborator’s website at: https://toniq.c2n.universite-paris-saclay.fr/fr/activites/smila/neuromimetic-photonics/

An example of code can be found out: http://github.com/SamiNakouzi/YAMADA

A good reference is found in [3]

[3] V. A. Pammi, K. Alfaro-Bittner, M. G. Clerc, S. Barbay, »Photonic computing with single and coupled spiking micropillar lasers » IEEE J. Sel. Top. Quantum Electron. 26, 1500307 (2020). https://doi.org/10.1109/JSTQE.2019.2929187