Exponential Transients in Continuous-Time Symmetric Hopfield Nets

0404166 - UIVT-O 20010035 RIV AT eng C - Conference Paper (international conference)
Šíma, Jiří - Orponen, P.
Exponential Transients in Continuous-Time Symmetric Hopfield Nets.
Artificial Neural Networks - ICANN 2001. Proceedings. Berlin: SpringerVerlag, 2001 - (Dorffner, G.; Bischof, H.; Hornik, K.), s. 806-814. Lecture Notes in Computer Science, 2130. ISBN 3-540-42486-5. ISSN 0302-9743.
[ICANN 2001 International Conference on Artificial Neural Networks /11./. Vienna (AT), 21.08.2001-25.08.2001]
R&D Projects: GA ČR GA201/01/1192; GA AV ČR IAB2030007
Institutional research plan: AV0Z1030915
Keywords : selforganization and dynamical systems * continuous-time Hopfield nets * Liapunov function * convergence time * combinatorial optimization
Subject RIV: BA - General Mathematics

We establish a fundamental result in the theory of continuosu-time neural computation, by showing that so called continuous-time symmetric Hopfield nets, whose asymptotic convergence is always guaranteed by the existence of a Liapunov function may, in the worst case, possess a transient period that is exponential in the network size. The result stands in contrast to e.g. the use of such network models in combinatorial optimization applications.
Permanent Link: http://hdl.handle.net/11104/0124433