Noninvertibility in neural networks

Ramiro Rico-Martinez, Yannis Kevrekidis, Raymond A. Adomaitis

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

We present a method for assessing certain validity aspects of predictions made by neural networks used to approximate continuous (in time) dynamical systems. This method searches for noninvertibility (non-uniqueness of the reverse time dynamics) of the fitted model, an indication of breakdown of 'proper' dynamical behavior. It is useful for computing bounds on the valid range of network predictions.

Original languageEnglish (US)
Title of host publication1993 IEEE International Conference on Neural Networks
Editors Anon
PublisherPubl by IEEE
Pages382-386
Number of pages5
ISBN (Print)0780312007
StatePublished - Jan 1 1993
Event1993 IEEE International Conference on Neural Networks - San Francisco, CA, USA
Duration: Mar 28 1993Apr 1 1993

Publication series

Name1993 IEEE International Conference on Neural Networks

Other

Other1993 IEEE International Conference on Neural Networks
CitySan Francisco, CA, USA
Period3/28/934/1/93

All Science Journal Classification (ASJC) codes

  • Engineering(all)

Fingerprint Dive into the research topics of 'Noninvertibility in neural networks'. Together they form a unique fingerprint.

Cite this