Summary
2023
Session Number:B3L-1
Session:
Number:B3L-14
Exploring Neural Cellular Automata for Simulating Cellular Interactions and Synchronization Phenomena
Gericke Eric, Schüle Martin,
pp.297-298
Publication Date:2023-09-21
Online ISSN:2188-5079
DOI:10.34385/proc.76.B3L-14
PDF download (204.2KB)
Summary:
Neural cellular automata (NCA) are a novel tool to learn and simulate the behaviour of interacting cells or agents. As a kind of convolutional network with residual blocks, NCA can learn local dynamics given some target pattern or behaviour. We demonstrate the versatility of this architecture on a variety of synchronization phenomena. In particular we implement a 1D and 2D local Kuramoto model in the NCA framework and analyze its computational performance in space and time. We also highlight additional potential applications.