Hybrid Learning of Regularization Neural Networks

Autoři: RNDr. Petra Vidnerová, Ph.D.
Mgr. Roman Neruda, Ph.D.
Rok vydání: 2010
Typ publikace: Konferenční příspěvek (mezinárodní konf.)
Periodikum: Artificial Intelligence and Soft Computing
Projekty: Algoritmy učení neuronových sítí založené na teorii regularizace
Bibliografická citace: VIDNEROVÁ, Petra, NERUDA, Roman. Hybrid Learning of Regularization Neural Networks. In: RUTKOWSKI, L., SCHERER, R., TADEUSIEWICZ, R., ZADEH, L.A., ZURADA, J.M., eds. Artificial Intelligence and Soft Computing. Vol. 2. Berlin: Springer, 2010, S. 124-131. Lecture Notes in Artificial Intelligence, 6114. ISBN 978-3-642-13231-5. ISSN 0302-9743
Dostupné z: doi:10.1007/978-3-642-13232-2_15
DOI: 10.1007/978-3-642-13232-2_15
ID: 345012

Abstrakt

Regularization theory presents a sound framework to solving supervised learning problems. However, the regularization networks have a large size corresponding to the size of training data. In this work we study a relationship between network complexity, i.e. number of hidden units, and approximation and generalization ability. We propose an incremental hybrid learning algorithm that produces smaller networks with performance similar to original regularization networks.