Perturbation Corrections in Approximate Inference: Mixture Modelling Applications Export

@article{citeulike:6048919, abstract = {Bayesian inference is intractable for many interesting models, making deterministic algorithms for approximate inference highly desirable. Unlike stochastic methods, which are exact in the limit, the accuracy of these approaches cannot be reasonably judged. In this paper we show how low order perturbation corrections to an expectation-consistent (EC) approximation can provide the necessary tools to ameliorate inference accuracy, and to give an indication of the quality of approximation without having to resort to Monte Carlo methods. Further comparisons are given with variational Bayes and parallel tempering (PT) combined with thermodynamic integration on a Gaussian mixture model. To obtain practical results we further generalize PT to temper from arbitrary distributions rather than a prior in Bayesian inference.}, author = {Paquet, Ulrich and Winther, Ole and Opper, Manfred}, citeulike-article-id = {6048919}, citeulike-linkout-0 = {http://www.jmlr.org/papers/volume10/paquet09a/paquet09a.pdf}, journal = {Journal of Machine Learning Research}, keywords = {bayesian, mixture, perturbation, statistics}, month = {June}, pages = {1263--1304}, posted-at = {2009-10-31 15:40:49}, priority = {2}, title = {Perturbation Corrections in Approximate Inference: Mixture Modelling Applications}, url = {http://www.jmlr.org/papers/volume10/paquet09a/paquet09a.pdf}, volume = {10}, year = {2009} }

TY - JOUR ID - citeulike:6048919 L3 - citeulike-article-id:6048919 N2 - Bayesian inference is intractable for many interesting models, making deterministic algorithms for approximate inference highly desirable. Unlike stochastic methods, which are exact in the limit, the accuracy of these approaches cannot be reasonably judged. In this paper we show how low order perturbation corrections to an expectation-consistent (EC) approximation can provide the necessary tools to ameliorate inference accuracy, and to give an indication of the quality of approximation without having to resort to Monte Carlo methods. Further comparisons are given with variational Bayes and parallel tempering (PT) combined with thermodynamic integration on a Gaussian mixture model. To obtain practical results we further generalize PT to temper from arbitrary distributions rather than a prior in Bayesian inference. JF - Journal of Machine Learning Research EP - 1304 TI - Perturbation Corrections in Approximate Inference: Mixture Modelling Applications VL - 10 SP - 1263 KW - bayesian KW - mixture KW - perturbation KW - statistics AU - Paquet, Ulrich AU - Winther, Ole AU - Opper, Manfred PY - 2009/06// UR - http://www.jmlr.org/papers/volume10/paquet09a/paquet09a.pdf ER -