Quantifying the accuracy of approximate diffusions and Markov chains
نویسندگان
چکیده
Markov chains and diffusion processes are indispensable tools in machine learning and statistics that are used for inference, sampling, and modeling. With the growth of large-scale datasets, the computational cost associated with simulating these stochastic processes can be considerable, and many algorithms have been proposed to approximate the underlying Markov chain or diffusion. A fundamental question is how the computational savings trade off against the statistical error incurred due to approximations. This paper develops general results that address this question. We bound the Wasserstein distance between the equilibrium distributions of two diffusions as a function of their mixing rates and the deviation in their drifts. We show that this error bound is tight in simple Gaussian settings. Our general result on continuous diffusions can be discretized to provide insights into the computational–statistical trade-off of Markov chains. As an illustration, we apply our framework to derive finite-sample error bounds of approximate unadjusted Langevin dynamics. We characterize computationconstrained settings where, by using fastto-compute approximate gradients in the Langevin dynamics, we obtain more accurate samples compared to using the exact gradients. Finally, as an additional application of our approach, we quantify the accuracy of approximate zig-zag sampling. Our theoretical analyses are supported by simulation experiments. Proceedings of the 20 International Conference on Artificial Intelligence and Statistics (AISTATS) 2017, Fort Lauderdale, Florida, USA. JMLR: W&CP volume 54. Copyright 2017 by the author(s).
منابع مشابه
Supplementary Material for Quantifying the Accuracy of Approximate Diffusions and Markov Chains
متن کامل
Markov chain approximations to non-symmetric diffusions with bounded coefficients
We consider a certain class of non-symmetric Markov chains and obtain heat kernel bounds and parabolic Harnack inequalities. Using the heat kernel estimates, we establish a sufficient condition for the family of Markov chains to converge to non-symmetric diffusions. As an application, we approximate non-symmetric diffusions in divergence form with bounded coefficients by non-symmetric Markov ch...
متن کاملConvergence Rates of Markov Chain Approximation Methods for Controlled Regime-Switching Diffusions with Stopping
This work summarizes our recent work on rates of convergence of Markov chain approximation methods for controlled switching diffusions, in which both continuous dynamics and discrete events coexist. The discrete events are formulated by continuous-time Markov chains to delineate random environment and other random factors that cannot be represented by diffusion processes. The cost function is o...
متن کاملToroidal diffusions and protein structure evolution
This chapter shows how toroidal diffusions are convenient methodological tools for modelling protein evolution in a probabilistic framework. The chapter addresses the construction of ergodic diffusions with stationary distributions equal to well-known directional distributions, which can be regarded as toroidal analogues of the Ornstein–Uhlenbeck process. The important challenges that arise in ...
متن کاملEmpirical Bayes Estimation in Nonstationary Markov chains
Estimation procedures for nonstationary Markov chains appear to be relatively sparse. This work introduces empirical Bayes estimators for the transition probability matrix of a finite nonstationary Markov chain. The data are assumed to be of a panel study type in which each data set consists of a sequence of observations on N>=2 independent and identically dis...
متن کامل