Global Asymptotic Stability and Stabilization of Long Short-Term Memory Neural Networks with Constant Weights and Biases

Shankar A. Deka*, Dušan M. Stipanović, Boris Murmann, Claire J. Tomlin

*Corresponding author for this work

Research output: Contribution to journalArticleScientificpeer-review

4 Citations (Scopus)

Abstract

In this paper, a global asymptotic stability condition for Long Short-Term Memory neural networks is presented. Since this condition is formulated in terms of the networks’ weight matrices and biases that are essentially control variables, the same condition can be viewed as a way to globally asymptotically stabilize these networks. The condition and how to compute numerical values for the weight matrices and biases are illustrated by a number of numerical examples.

Original languageEnglish
Pages (from-to)231-243
Number of pages13
JournalJournal of Optimization Theory and Applications
Volume181
Issue number1
DOIs
Publication statusPublished - 15 Apr 2019
MoE publication typeA1 Journal article-refereed

Keywords

  • Global asymptotic stability
  • Neural networks
  • Stabilization

Fingerprint

Dive into the research topics of 'Global Asymptotic Stability and Stabilization of Long Short-Term Memory Neural Networks with Constant Weights and Biases'. Together they form a unique fingerprint.

Cite this