Saved in:
Bibliographic Details
Main Authors: Brekelmans, Rob, Masrani, Vaden, Bui, Thang, Wood, Frank, Galstyan, Aram, Steeg, Greg Ver, Nielsen, Frank
Format: Preprint
Published: 2020
Subjects:
Online Access:https://arxiv.org/abs/2012.07823
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911853416808448
author Brekelmans, Rob
Masrani, Vaden
Bui, Thang
Wood, Frank
Galstyan, Aram
Steeg, Greg Ver
Nielsen, Frank
author_facet Brekelmans, Rob
Masrani, Vaden
Bui, Thang
Wood, Frank
Galstyan, Aram
Steeg, Greg Ver
Nielsen, Frank
contents Annealed importance sampling (AIS) is the gold standard for estimating partition functions or marginal likelihoods, corresponding to importance sampling over a path of distributions between a tractable base and an unnormalized target. While AIS yields an unbiased estimator for any path, existing literature has been primarily limited to the geometric mixture or moment-averaged paths associated with the exponential family and KL divergence. We explore AIS using $q$-paths, which include the geometric path as a special case and are related to the homogeneous power mean, deformed exponential family, and $α$-divergence.
format Preprint
id arxiv_https___arxiv_org_abs_2012_07823
institution arXiv
publishDate 2020
record_format arxiv
spellingShingle Annealed Importance Sampling with q-Paths
Brekelmans, Rob
Masrani, Vaden
Bui, Thang
Wood, Frank
Galstyan, Aram
Steeg, Greg Ver
Nielsen, Frank
Machine Learning
Annealed importance sampling (AIS) is the gold standard for estimating partition functions or marginal likelihoods, corresponding to importance sampling over a path of distributions between a tractable base and an unnormalized target. While AIS yields an unbiased estimator for any path, existing literature has been primarily limited to the geometric mixture or moment-averaged paths associated with the exponential family and KL divergence. We explore AIS using $q$-paths, which include the geometric path as a special case and are related to the homogeneous power mean, deformed exponential family, and $α$-divergence.
title Annealed Importance Sampling with q-Paths
topic Machine Learning
url https://arxiv.org/abs/2012.07823