Learning of correlated patterns by simple perceptrons - Condensed Matter > Disordered Systems and Neural NetworksReportar como inadecuado




Learning of correlated patterns by simple perceptrons - Condensed Matter > Disordered Systems and Neural Networks - Descarga este documento en PDF. Documentación en PDF para descargar gratis. Disponible también para leer online.

Abstract: Learning behavior of simple perceptrons is analyzed for a teacher-studentscenario in which output labels are provided by a teacher network for a set ofpossibly correlated input patterns, and such that teacher and student networksare of the same type. Our main concern is the effect of statisticalcorrelations among the input patterns on learning performance. For thispurpose, we extend to the teacher-student scenario a methodology for analyzingrandomly labeled patterns recently developed in {\em J. Phys. A: Math. Theor.}{\bf 41}, 324013 2008. This methodology is used for analyzing situations inwhich orthogonality of the input patterns is enhanced in order to optimize thelearning performance.



Autor: Takashi Shinzato, Yoshiyuki Kabashima

Fuente: https://arxiv.org/







Documentos relacionados