MEMLJun 28, 2019

Learning Markov models via low-rank optimization

arXiv:1907.00113v220 citations
Originality Incremental advance
AI Analysis

This addresses the problem of efficient system modeling from limited data for optimization and decision-making, with incremental improvements in estimation methods.

The paper tackles learning a Markov model from a single trajectory when the transition model has low rank, showing that accurate estimation is possible with a trajectory length proportional to the total number of states. It proposes convex and nonconvex maximum likelihood estimators, deriving statistical rates and establishing a minimax lower bound, with empirical results showing the nonconvex estimator's superiority.

Modeling unknown systems from data is a precursor of system optimization and sequential decision making. In this paper, we focus on learning a Markov model from a single trajectory of states. Suppose that the transition model has a small rank despite of having a large state space, meaning that the system admits a low-dimensional latent structure. We show that one can estimate the full transition model accurately using a trajectory of length that is proportional to the total number of states. We propose two maximum likelihood estimation methods: a convex approach with nuclear-norm regularization and a nonconvex approach with rank constraint. We explicitly derive the statistical rates of both estimators in terms of the Kullback-Leiber divergence and the $\ell_2$ error and also establish a minimax lower bound to assess the tightness of these rates. For computing the nonconvex estimator, we develop a novel DC (difference of convex function) programming algorithm that starts with the convex M-estimator and then successively refines the solution till convergence. Empirical experiments demonstrate consistent superiority of the nonconvex estimator over the convex one.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes