Summary
International Symposium on Nonlinear Theory and Its Applications
2022
Session Number:A3L-D
Session:
Number:A3L-D-04
Visual Predictive Coding Model with Reservoir Computing for Reinforcement Learning Tasks in 3D Environment
Tomohito Izumi , Yuichi Katori,
pp.91-94
Publication Date:12/12/2022
Online ISSN:2188-5079
DOI:10.34385/proc.71.A3L-D-04
PDF download (2.5MB)
Summary:
Computational models are an indispensable tool for understanding the mechanism of the brain. Previous studies employ reservoir computing to construct a predictive coding model and have shown its ability to replicate the brain's properties. However, reservoir computing models cannot be directly applied to broad brain functions due to performance limitations. Here, we propose a visual predictive coding model with reservoir computing that can handle the high-dimensional input to extend the scope of the application. We believe that our approach presents a novel dynamical mechanism of visual processing in the brain and fundamental technology for a brain-like artificial intelligence system.