Cellular automata as convolutional neural networks

09/09/2018
by   William Gilpin, et al.
6

Deep learning techniques have recently demonstrated broad success in predicting complex dynamical systems ranging from turbulence to human speech, motivating broader questions about how neural networks encode and represent dynamical rules. We explore this problem in the context of cellular automata (CA), simple dynamical systems that are intrinsically discrete and thus difficult to analyze using standard tools from dynamical systems theory. We show that any CA may readily be represented using a convolutional neural network with a network-in-network architecture. This motivates our development of a general convolutional multilayer perceptron architecture, which we find can learn the dynamical rules for arbitrary CA when given videos of the CA as training data. In the limit of large network widths, we find that training dynamics are strongly stereotyped across replicates, and that common patterns emerge in the structure of networks trained on different CA rulesets. We train ensembles of networks on randomly-sampled CA, and we probe how the trained networks internally represent the CA rules using an information-theoretic technique based on distributions of layer activation patterns. We find that CA with simpler rule tables produce trained networks with hierarchical structure and layer specialization, while more complex CA tend to produce shallower representations---illustrating how the underlying complexity of the CA's rules influences the specificity of these internal representations. Our results suggest how the entropy of a physical process can affect its representation when learned by neural networks.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
07/03/2019

A general representation of dynamical systems for reservoir computing

Dynamical systems are capable of performing computation in a reservoir c...
research
05/02/2023

Diddy: a Python toolbox for infinite discrete dynamical systems

We introduce Diddy, a collection of Python scripts for analyzing infinit...
research
05/13/2022

Physics guided neural networks for modelling of non-linear dynamics

The success of the current wave of artificial intelligence can be partly...
research
04/18/2012

Discrete Dynamical Genetic Programming in XCS

A number of representation schemes have been presented for use within Le...
research
01/25/2023

E(n)-equivariant Graph Neural Cellular Automata

Cellular automata (CAs) are computational models exhibiting rich dynamic...
research
01/26/2012

Discrete and fuzzy dynamical genetic programming in the XCSF learning classifier system

A number of representation schemes have been presented for use within le...
research
05/14/2020

Echo State Networks trained by Tikhonov least squares are L2(μ) approximators of ergodic dynamical systems

Echo State Networks (ESNs) are a class of single-layer recurrent neural ...

Please sign up or login with your details

Forgot password? Click here to reset