We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:

References & Citations


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Mathematics > Statistics Theory

Title: Quantifying the accuracy of approximate diffusions and Markov chains

Abstract: 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 computation-constrained settings where, by using fast-to-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.
Comments: In Proceedings of the 19th International Conference on Artificial Intelligence and Statistics (AISTATS 2017)
Subjects: Statistics Theory (math.ST); Probability (math.PR); Computation (stat.CO); Machine Learning (stat.ML)
Cite as: arXiv:1605.06420 [math.ST]
  (or arXiv:1605.06420v4 [math.ST] for this version)

Submission history

From: Jonathan Huggins [view email]
[v1] Fri, 20 May 2016 16:17:22 GMT (44kb,D)
[v2] Wed, 12 Oct 2016 17:30:56 GMT (56kb,D)
[v3] Wed, 1 Mar 2017 15:06:21 GMT (58kb,D)
[v4] Wed, 30 Aug 2017 14:50:01 GMT (58kb,D)

Link back to: arXiv, form interface, contact.