Abstract:We consider the problem of learning graphical models, also known as Markov random fields (MRFs) from temporally correlated samples. As in many traditional statistical settings, fundamental results in the area all assume independent samples from the distribution. However, these samples generally will not directly correspond to more realistic observations from nature, which instead evolve according to some stochastic process. From the computational lens, even generating a single sample from the true MRF distribution is intractable unless $\mathsf{NP}=\mathsf{RP}$, and moreover, any algorithm to learn from i.i.d. samples requires prohibitive runtime due to hardness reductions to the parity with noise problem. These computational barriers for sampling and learning from the i.i.d. setting severely lessen the utility of these breakthrough results for this important task; however, dropping this assumption typically only introduces further algorithmic and statistical complexities. In this work, we surprisingly demonstrate that the direct trajectory data from a natural evolution of the MRF overcomes the fundamental computational lower bounds to efficient learning. In particular, we show that given a trajectory with $\widetilde{O}_k(n)$ site updates of an order $k$ MRF from the Glauber dynamics, a well-studied, natural stochastic process on graphical models, there is an algorithm that recovers the graph and the parameters in $\widetilde{O}_k(n^2)$ time. By contrast, all prior algorithms for learning order $k$ MRFs inherently suffer from $n^{\Theta(k)}$ runtime even in sparse instances due to the reductions to sparse parity with noise. Our results thus surprisingly show that this more realistic, but intuitively less tractable, model for MRFs actually leads to efficiency far beyond what is known and believed to be true in the traditional i.i.d. case.

Geometric ergodicity of Rao and Teh's algorithm for Markov jump processes

Using Stein’s Method to Analyze Euler–Maruyama Approximations of Regime-Switching Jump Diffusion Processes

Variable-Step Euler–Maruyama Approximations of Regime-Switching Jump Diffusion Processes

Geometric ergodicity of trans-dimensional Markov chain Monte Carlo algorithms

On the Geometric Ergodicity of Metropolis-Hastings Algorithms for Lattice Gaussian Sampling

Geometric allocation approaches in Markov chain Monte Carlo

Geometric Ergodicity and the Spectral Gap of Non-Reversible Markov Chains

Estimating the Mixing Coefficients of Geometrically Ergodic Markov Processes

Exponential ergodicity of some Markov dynamical system with application to a Poisson driven stochastic differential equation

Ergodicity and first passage probability of regime-switching geometric Brownian motions

On the $α$-lazy version of Markov chains in estimation and testing problems

Markov chain Monte Carlo Methods For Lattice Gaussian Sampling:Convergence Analysis and Enhancement

Equivalences of Geometric Ergodicity of Markov Chains

Efficient Sampling on Riemannian Manifolds via Langevin MCMC

Information-geometric Markov Chain Monte Carlo methods using Diffusions

Error Bounds for Augmented Truncations of Discrete-Time Block-Monotone Markov Chains under Geometric Drift Conditions

On sampling graphical Markov models

Geometry and Dynamics for Markov Chain Monte Carlo

Bypassing the Noisy Parity Barrier: Learning Higher-Order Markov Random Fields from Dynamics

Unbiased approximation of the ergodic measure for piecewise $α$-stable Ornstein-Uhlenbeck processes arising in queueing networks

A stochastic algorithm approach for the elephant random walk with applications