Online learning and optimization of Markov jump linear models

Abstract

The problem of online learning and optimization of unknown Markov jump linear models is considered. A new online learning algorithm, referred to as Markovian simultaneous perturbations stochastic approximation (MSPSA), is proposed. It is shown that ν/ MSPSA achieves the minimax regret order of Θ(√T). Using the Van Trees inequality… (More)
DOI: 10.1109/ICASSP.2016.7472085

Topics

1 Figure or Table

Slides referencing similar topics