Alpha-Beta Cellular Automata

Abstract

Alpha-Beta Cellular Automata arise as a conceptual bridge between two research areas which have remained disjoint to this day: Alpha-Beta associative memories on one hand and cellular automata on the other hand. The results presented in this work make it possible to apply tools developed on the basis of the Alpha-Beta associative models to the field of cellular automata and vice versa: one can use the theoretical body of cellular automata in applications of the Alpha-Beta associative models. Specifically, the proposed model is applied to the task of classification on the MNIST database of handwritten digits as published by the US National Institute of Standards and Technology (NIST). The results are competitive when compared to those given by other algorithms applied to the same database.