Accuracy Estimates for Surrogate Solutions of Integral Equations by Neural Networks

Autoři: RNDr. Věra Kůrková, DSc.
Rok vydání: 2012
Typ publikace: Konferenční příspěvek (mezinárodní konf.)
Periodikum: SOFSEM 2012: Theory and Practice of Computer Science
Projekty: Analýza inteligentních distribuovaných výpočetních systémů
Bibliografická citace: KŮRKOVÁ, Věra. Accuracy Estimates for Surrogate Solutions of Integral Equations by Neural Networks. In: BIELIKOVÁ, M., FRIEDRICH, G., GOTTLOB, G., KATZENBEISSER, S., ŠPÁNEK, R., TURÁN, G., eds. SOFSEM 2012: Theory and Practice of Computer Science. Vol. 2. Prague: Institute of Computer Science AS CR, 2012, S. 95-102. ISBN 978-80-87136-13-3.
ID: 374354

Abstrakt

Surrogate solutions of integral equations by neural networks are investigated theoretically. Upper bounds on speed of convergence of approximate solutions computable by neural networks with increasing model complexity to exact solutions described by Fredholm theorem are derived.