conftrace_

2014 ICML ICML 2014

Lower Bounds for the Gibbs Sampler over Mixtures of Gaussians

Abstract

The mixing time of a Markov chain is the minimum time t necessary for the total variation distance between the distribution of the Markov chain’s current state X_t and its stationary distribution to fall below some ε> 0. In this paper, we present lower bounds for the mixing time of the Gibbs sampler over Gaussian mixture models with Dirichlet priors.

🌉 Interdisciplinary Bridge - Artificial Intelligence and Mathematics & Optimization

🐣 Hot Topic Early Bird - gaussian mixture model

Authors

Christopher Tosh, Sanjoy Dasgupta

Topics

Machine Learning > Optimization & Theory > Bayesian Inference Mathematics & Optimization > Mathematics > Probability Mathematics & Optimization > Optimization > Stochastic Methods Mathematics & Optimization > Probability > Stochastic Processes

Keywords

bayesian inference markov chain monte carlo gibbs sampling markov chain stochastic process gibbs sampler mixing time gaussian mixture model mixture model dirichlet prior

More from ICML 2014

All ICML 2014 papers →

Learning Modular Structures from Network Data and Node Variables

Probabilistic Partial Canonical Correlation Analysis

Skip Context Tree Switching

Marginalized Denoising Auto-encoders for Nonlinear Representations

Gaussian Processes for Bayesian Estimation in Ordinary Differential Equations

Fast Multi-stage Submodular Maximization