Collective and synchronous dynamics of photonic spiking neurons

Takahiro Inagaki,Hiroki Takesue,Takuya Ikuta,Ryoichi Kasahara,Kensuke Inaba,Toshimori Honjo,Timothée Leleu,Koji Enbutsu,Kazuyuki Aihara,Takeshi Umeki

doi:10.1038/s41467-021-22576-4

Takahiro Inagaki, Hiroki Takesue + Show 8 more

Open Access

https://doi.org/10.1038/s41467-021-22576-4

Copy DOI

Abstract

Nonlinear dynamics of spiking neural networks have recently attracted much interest as an approach to understand possible information processing in the brain and apply it to artificial intelligence. Since information can be processed by collective spiking dynamics of neurons, the fine control of spiking dynamics is desirable for neuromorphic devices. Here we show that photonic spiking neurons implemented with paired nonlinear optical oscillators can be controlled to generate two modes of bio-realistic spiking dynamics by changing optical-pump amplitude. When the photonic neurons are coupled in a network, the interaction between them induces an effective change in the pump amplitude depending on the order parameter that characterizes synchronization. The experimental results show that the effective change causes spontaneous modification of the spiking modes and firing rates of clustered neurons, and such collective dynamics can be utilized to realize efficient heuristics for solving NP-hard combinatorial optimization problems.

Highlights

Nonlinear dynamics of spiking neural networks have recently attracted much interest as an approach to understand possible information processing in the brain and apply it to artificial intelligence
It was shown that some neuron models can generate both class-I and class-II spiking modes depending on the values of the model parameters[27,28,29,30,31,32], the spiking modes of component neurons in spiking neural networks are usually fixed in advance
A photonic spiking neuronal network was developed by utilizing a network of degenerate optical parametric oscillators (DOPOs) pulses in a fiber-ring cavity, which has been used for simulating an Ising spin network and solving combinatorial optimization problems[33,34,35,36,37]

Summary

Introduction

Nonlinear dynamics of spiking neural networks have recently attracted much interest as an approach to understand possible information processing in the brain and apply it to artificial intelligence. We show that photonic spiking neurons implemented with paired nonlinear optical oscillators can be controlled to generate two modes of bio-realistic spiking dynamics by changing optical-pump amplitude. Various nonlinear dynamics of coupled parametric oscillators, which can be utilized for artificial spiking neurons, have been proposed[23,24], and an effect of collective dynamics on the synchronization of the spikes has been demonstrated with an array of vertical cavity surface emitting lasers[25]. The spiking mode of our proposed photonic neuron, on the other hand, can change due to collective and synchronous dynamics of the network for spontaneous information processing because the spiking dynamics can be controlled by tuning optical-pump amplitudes of the DOPOs. Network experiments with 240 DOPO neurons revealed that input signals from the correlating neurons can induce an effective change in the pump amplitude. The experimental results showed that the self-tuning effect of collective spiking dynamics can be utilized for solving combinatorial optimization problems by using methods related to self-organized criticality

Methods

Results

Conclusion