A Neural Network Architecture which uses Cellular Automata to get Context-Specific Associative Memory

Abstract In our model we use a Neural Cellular Automata model (NCA) for the generation of a context-specific associative memory. Based on a short structural description of the two-dimensional NCA with its corresponding equations in the following part, we demonstrate how our model architecture in which a special kind of NCA and a common associative memory interact is capable of extracting the inherent structure of an arbitrary set of input data.