Asynchronous Cellular Automata and Pattern Classification

TL;DR

This paper introduces a novel pattern classification method using asynchronous cellular automata (ACAs), identifying effective ACAs with fixed point convergence properties and demonstrating superior performance over existing algorithms.

Contribution

The paper proposes a systematic approach to select ACAs for pattern classification based on fixed point convergence criteria and provides experimental validation of their effectiveness.

Findings

Identified 71 effective ACAs out of 256 for pattern classification.

Effective ACAs outperform many standard algorithms in experiments.

Proposed method offers a reliable and efficient classification approach.

Abstract

This paper designs an efficient two-class pattern classifier utilizing asynchronous cellular automata (ACAs). The two-state three-neighborhood one-dimensional ACAs that converge to fixed points from arbitrary seeds are used here for pattern classification. To design the classifier, we first identify a set of ACAs that always converge to fixed points from any seeds with following properties - (1) each ACA should have at least two but not huge number of fixed point attractors, and (2) the convergence time of these ACAs are not to be exponential. In order to address the first issue, we propose a graph, coined as fixed point graph of an ACA that facilitates in counting the fixed points. We further perform an experimental study to estimate the convergence time of ACAs, and find that there are some convergent ACAs which demand exponential convergence time. Finally, we find that there are 71 (out of 256) ACAs which can be effective candidates as pattern classifier. We use each of the candidate ACAs on some standard data sets, and observe the effectiveness of each ACAs as pattern classifier. It is observed that the proposed classifier is very competitive and performs reliably better than many standard existing algorithms.