USING CHAOS TO GENERATE KEYS FOR ASSOCIATIVE NOISE-LIKE CODING MEMORIES

In the noise-like coding model of associative memory, the core of the information-coding mechanism lies in the key-production process, by wh ich a given pattern is transformed into a corresponding noise-like key for both information storage and retrieval. In this paper, it is show n how the pseudorandom behaviour typical of chaotic processes can be e xploited to obtain noise-like patterns from supplied input patterns by a deterministic procedure. A mathematical analysis is carried out to prove the validity of the approach: It demonstrates that the generated patterns satisfy the noise-like constraints imposed by the associativ e model on keys, The main advantage is the possibility of performing c omputations at the local level. Experimental results confirm both the correctness of the theoretical derivations and the effectiveness of th e proposed methodology in (visual) pattern classification applications .