Variational MCMC

01/10/2013
by   Nando de Freitas, et al.
0

We propose a new class of learning algorithms that combines variational approximation and Markov chain Monte Carlo (MCMC) simulation. Naive algorithms that use the variational approximation as proposal distribution can perform poorly because this approximation tends to underestimate the true variance and other features of the data. We solve this problem by introducing more sophisticated MCMC algorithms. One of these algorithms is a mixture of two MCMC kernels: a random walk Metropolis kernel and a blockMetropolis-Hastings (MH) kernel with a variational approximation as proposaldistribution. The MH kernel allows one to locate regions of high probability efficiently. The Metropolis kernel allows us to explore the vicinity of these regions. This algorithm outperforms variationalapproximations because it yields slightly better estimates of the mean and considerably better estimates of higher moments, such as covariances. It also outperforms standard MCMC algorithms because it locates theregions of high probability quickly, thus speeding up convergence. We demonstrate this algorithm on the problem of Bayesian parameter estimation for logistic (sigmoid) belief networks.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
05/31/2021

Variational Combinatorial Sequential Monte Carlo Methods for Bayesian Phylogenetic Inference

Bayesian phylogenetic inference is often conducted via local or sequenti...
research
03/07/2022

Convergence Speed and Approximation Accuracy of Numerical MCMC

When implementing Markov Chain Monte Carlo (MCMC) algorithms, perturbati...
research
04/24/2020

Increasing the efficiency of Sequential Monte Carlo samplers through the use of approximately optimal L-kernels

By facilitating the generation of samples from arbitrary probability dis...
research
11/20/2019

Ensuring Reliable Monte Carlo Estimates of Network Properties

The literature in social network analysis has largely focused on methods...
research
12/31/2020

Nonreversible MCMC from conditional invertible transforms: a complete recipe with convergence guarantees

Markov Chain Monte Carlo (MCMC) is a class of algorithms to sample compl...
research
12/12/2012

Decayed MCMC Filtering

Filtering---estimating the state of a partially observable Markov proces...
research
01/18/2017

Converting Cascade-Correlation Neural Nets into Probabilistic Generative Models

Humans are not only adept in recognizing what class an input instance be...

Please sign up or login with your details

Forgot password? Click here to reset