Nonlinear ICA through low-complexity autoencoders

Sepp Hochreiter, Juergen Schmidhuber

Research output: Contribution to journalArticlepeer-review

2 Scopus citations

Abstract

Flat Minimum Search, a regularizer algorithm for finding low-complexity networks describable by few bits of information, is employed to train autocoders. Flat minima are regions in weight space where there is tolerable small error and the weights can be perturbed without greatly affecting the network's output. The procedure reveals an important connection between regularization and independent component analysis. This connection may represent a first step towards unification of regularization and unsupervised learning.
Original languageEnglish (US)
JournalProceedings - IEEE International Symposium on Circuits and Systems
Volume5
StatePublished - Jan 1 1999
Externally publishedYes

Bibliographical note

Generated from Scopus record by KAUST IRTS on 2022-09-14

Fingerprint

Dive into the research topics of 'Nonlinear ICA through low-complexity autoencoders'. Together they form a unique fingerprint.

Cite this