Robust Implementation of Finite Automata by Recurrent RBF Networks

0403972 - UIVT-O 20000112 RIV DE eng C - Conference Paper (international conference)
Šorel, Michal - Šíma, Jiří
Robust Implementation of Finite Automata by Recurrent RBF Networks.
SOFSEM'2000: Theory and Practice of Informatics. Berlin: Springer, 2000 - (Hlaváč, V.; Jeffery, K.; Wiedermann, J.), s. 431-439. Lecture Notes in Computer Science, 1963. ISBN 3-540-41348-0. ISSN 0302-9743.
[SOFSEM 2000 /27./. Milovy (CZ), 25.11.2000-01.12.2000]
R&D Projects: GA AV ČR IAB2030007
Institutional research plan: AV0Z1030915
Subject RIV: BA - General Mathematics

In this paper a recurrent network, which consists of O(sqrt(m log m)) RBF (radial basis functions) units with maximum norm employing any activation function that has different values in at least two nonnegative points, is constructed so as to implement a given deterministic finite automaton with m states. The underlying simulation proves to be robust with respect to analog noise for a large class of smooth activation functions with a special type of inflexion.
Permanent Link: http://hdl.handle.net/11104/0124256