Brief Paper Associative memory design using overlapping decompositions q

This paper discusses the use of decomposition techniques in the design of associative memories via arti"cial neural networks. In particular, a disjoint decomposition which allows an independent design of lower-dimensional subnetworks and an overlapping decomposition which allows subnetworks to share common parts, are analyzed. It is shown by a simple example that overlapping decompositions may help in certain cases where design by disjoint decompositions fails. With this motivation, an algorithm is provided to synthesize neural networks using the concept of overlapping decompositions. Applications of the proposed design procedure to a benchmark example from the literature and to a pattern recognition problem indicate that it may improve the e!ectiveness of the existing methods. ( 2001 Published by Elsevier Science Ltd.

[1]  A.N. Michel,et al.  Analysis and synthesis of a class of discrete-time neural networks described on hypercubes , 1991, IEEE Trans. Neural Networks.

[2]  Stefen Hui,et al.  Synthesis of Brain-State-in-a-Box (BSB) based associative memories , 1994, IEEE Trans. Neural Networks.

[3]  D. Siljak,et al.  Decentralized control with overlapping information sets , 1981 .

[4]  Renzo Perfetti,et al.  A synthesis procedure for brain-state-in-a-box neural networks , 1995, IEEE Trans. Neural Networks.

[5]  I. Guyon,et al.  Information storage and retrieval in spin-glass like neural networks , 1985 .

[6]  Umit Ozguner,et al.  Decentralized parallel distributed compensator design for Takagi-Sugeno fuzzy systems , 1999, Proceedings of the 38th IEEE Conference on Decision and Control (Cat. No.99CH36304).

[7]  André Titli,et al.  Overlapping vs partitioning in block-iteration methods: Application in large-scale system theory , 1989, Autom..

[8]  Wolfgang Porod,et al.  Analysis and synthesis of a class of neural networks: variable structure systems with infinite grain , 1989 .

[9]  J J Hopfield,et al.  Neural networks and physical systems with emergent collective computational abilities. , 1982, Proceedings of the National Academy of Sciences of the United States of America.

[10]  David Mautner Himmelblau,et al.  Decomposition of large-scale problems , 1973 .

[11]  Ikeda,et al.  Overlapping decompositions, expansions, and contractions of dynamic systems , 1979 .

[12]  D. Siljak,et al.  An inclusion principle for dynamic systems , 1984 .