MODELING "COGNITION" WITH NONLINEAR DYNAMIC SYSTEMS

In this paper we consider realization of information processing and recognition with dynamic systems. A method for storing and processing information is described, which is connected with symmetries and in which digital information blocks are related to dynamic attractors (periodic orbits or chaotic attractors) of a specially designed nonlinear dynamic system. Such a system has interesting features concerning information processing, in particular, associative access (retrieval of the whole stored image when a small part of it is given as request), search by content, novelty filter, etc. The designed dynamic systems can be used as storage for texts, pictures, digital sequences, etc. Practical examples of using this approach to create information search engines, information archives with associative access to the stored data, and other data management solutions are given.