Internal dynamics of recurrent neural networks trained to generate complex spatiotemporal patterns

Oleg V Maslennikov; Chao Gao; Vladimir I Nekorkin

doi:10.1063/5.0166359

Internal dynamics of recurrent neural networks trained to generate complex spatiotemporal patterns

Chaos. 2023 Sep 1;33(9):093125. doi: 10.1063/5.0166359.

Authors

Oleg V Maslennikov¹, Chao Gao², Vladimir I Nekorkin¹

Affiliations

¹ Federal Research Center A.V. Gaponov-Grekhov Institute of Applied Physics of the Russian Academy of Sciences, Nizhny Novgorod, Russia.
² School of Artificial Intelligence, Optics and Electronics, Northwestern Polytechnical University, Xian, China.

PMID: 37722673
DOI: 10.1063/5.0166359

Abstract

How complex patterns generated by neural systems are represented in individual neuronal activity is an essential problem in computational neuroscience as well as machine learning communities. Here, based on recurrent neural networks in the form of feedback reservoir computers, we show microscopic features resulting in generating spatiotemporal patterns including multicluster and chimera states. We show the effect of individual neural trajectories as well as whole-network activity distributions on exhibiting particular regimes. In addition, we address the question how trained output weights contribute to the autonomous multidimensional dynamics.