Learning Mixtures of Linear Dynamical Systems
Yanxi Chen · H. Vincent Poor
Keywords:
T: Learning Theory
MISC: Supervised Learning
MISC: Transfer, Multitask and Meta-learning
MISC: Unsupervised and Semi-supervised Learning
T: Miscellaneous Aspects of Machine Learning
MISC: Sequential, Network, and Time Series Modeling
OPT: Control and Optimization
PM: Spectral Methods
Theory
2022 Poster
Abstract
We study the problem of learning a mixture of multiple linear dynamical systems (LDSs) from unlabeled short sample trajectories, each generated by one of the LDS models. Despite the wide applicability of mixture models for time-series data, learning algorithms that come with end-to-end performance guarantees are largely absent from existing literature. There are multiple sources of technical challenges, including but not limited to (1) the presence of latent variables (i.e. the unknown labels of trajectories); (2) the possibility that the sample trajectories might have lengths much smaller than the dimension $d$ of the LDS models; and (3) the complicated temporal dependence inherent to time-series data. To tackle these challenges, we develop a two-stage meta-algorithm, which is guaranteed to efficiently recover each ground-truth LDS model up to error $\tilde{O}(\sqrt{d/T})$, where $T$ is the total sample size. We validate our theoretical studies with numerical experiments, confirming the efficacy of the proposed algorithm.
Chat is not available.
Successful Page Load