Implicitly Adaptive Importance Sampling

Topi Paananen, Juho Piironen, Paul-Christian Bürkner, Aki Vehtari
6/20/2019
stat.COstat.MEstat.ML

Official code: topipa/iter-mm-paper (9 stars, last commit 2021-05-13).

Abstract

Adaptive importance sampling is a class of techniques for finding good proposal distributions for importance sampling. Often the proposal distributions are standard probability distributions whose parameters are adapted based on the mismatch between the current proposal and a target distribution. In this work, we present an implicit adaptive importance sampling method that applies to complicated distributions which are not available in closed form. The method iteratively matches the moments of a set of Monte Carlo draws to weighted moments based on importance weights. We apply the method to Bayesian leave-one-out cross-validation and show that it performs better than many existing parametric adaptive importance sampling methods while being computationally inexpensive.

View on arXivView PDF

Code Implementations(1)

topipa/iter-mm-paperOfficial100%
90R, TeXJun 10, 20195 years ago

Similar Papers

Perturbative adaptive importance sampling for Bayesian LOO cross-validation

82% match1 repo12024

Sampling as Bandits: Evaluation-Efficient Design for Black-Box Densities

81% match2025

BR-SNIS: Bias Reduced Self-Normalized Importance Sampling

81% match2022

Online Variance Reduction for Stochastic Optimization

80% match1 repo152018

Enhanced Importance Sampling through Latent Space Exploration in Normalizing Flows

80% match1 repo12025

Pareto Smoothed Importance Sampling

79% match2 repos1592015

Cite this paper

@article{paananen2019implicitly,
  title  = {Implicitly Adaptive Importance Sampling},
  author = {Topi Paananen and Juho Piironen and Paul-Christian Bürkner and Aki Vehtari},
  year   = {2019},
  eprint = {1906.08850},
  archivePrefix = {arXiv},
  url    = {https://arxiv.org/abs/1906.08850}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/1906.08850.svg)](https://codeofpaper.com/paper/1906.08850)

Discussion