• Corpus ID: 221397170

# Adaptive Path Sampling in Metastable Posterior Distributions

@article{Yao2020AdaptivePS,
title={Adaptive Path Sampling in Metastable Posterior Distributions},
author={Yuling Yao and Collin Cademartori and Aki Vehtari and Andrew Gelman},
journal={arXiv: Computation},
year={2020}
}
• Published 1 September 2020
• Computer Science
• arXiv: Computation
The normalizing constant plays an important role in Bayesian computation, and there is a large literature on methods for computing or approximating normalizing constants that cannot be evaluated in closed form. When the normalizing constant varies by orders of magnitude, methods based on importance sampling can require many rounds of tuning. We present an improved approach using adaptive path sampling, iteratively reducing gaps between the base and target. Using this adaptive strategy, we…
4 Citations

## Figures from this paper

Continuously-Tempered PDMP Samplers
• Mathematics, Computer Science
• 2022
An extended distribution is introduced over the state of the posterior distribution and an inverse temperature, which interpolates between a tractable distribution when the inverse temperature is 0 and the posterior when the reverse temperature is 1, and it is shown how PDMPs, and particularly the Zig-Zag sampler, can be implemented to sample from such an extended distribution.
Density Ratio Estimation via Infinitesimal Classification
• Computer Science
AISTATS
• 2022
This work proposes DRE-8, a divide-and-conquer approach to reduce DRE to a series of easier subproblems, and shows that traditional (Stein) scores can be used to obtain integration paths that connect regions of high density in both distributions, improving performance in practice.
Bayesian Workflow.
• Computer Science
• 2020
The Bayesian approach to data analysis provides a powerful way to handle uncertainty in all observations, model parameters, and model structure using probability theory, and this work reviews all aspects of workflow in the context of several examples.
Stacking for Non-mixing Bayesian Computations: The Curse and Blessing of Multimodal Posteriors
• Computer Science
• 2020
Using parallel runs of MCMC, variational, or mode-based inference to hit as many modes or separated regions as possible, and then combining these using importance sampling based Bayesian stacking, a scalable method for constructing a weighted average of distributions so as to maximize cross-validated prediction utility is proposed.

## References

SHOWING 1-10 OF 53 REFERENCES
Annealed importance sampling
It is shown how one can use the Markov chain transitions for such an annealing sequence to define an importance sampler, which can be seen as a generalization of a recently-proposed variant of sequential importance sampling.
Rank-normalization, folding, and localization: An improved R-hat for assessing convergence of MCMC
• Computer Science, Mathematics
• 2020
A collection of quantile-based local efficiency measures, along with a practical approach for computing Monte Carlo error estimates for quantiles, and a suggestion that common trace plots should be replaced with rank plots from multiple chains.
Continuously Tempered Hamiltonian Monte Carlo
• Computer Science
UAI
• 2017
A method for augmenting the Hamiltonian system with an extra continuous temperature control variable which allows the dynamic to bridge between sampling a complex target distribution and a simpler unimodal base distribution and allows the normalisation constant of the target distribution to be estimated.
• Computer Science
ICML
• 2014
K Kernel Adaptive Metropolis-Hastings outperforms competing fixed and adaptive samplers on multivariate, highly nonlinear target distributions, arising in both real-world and synthetic examples.
The No-U-turn sampler: adaptively setting path lengths in Hamiltonian Monte Carlo
• Computer Science
J. Mach. Learn. Res.
• 2014
The No-U-Turn Sampler (NUTS), an extension to HMC that eliminates the need to set a number of steps L, and derives a method for adapting the step size parameter {\epsilon} on the fly based on primal-dual averaging.
Simulating Normalizing Constants: From Importance Sampling to Bridge Sampling to Path Sampling
• Computer Science
• 1998
It is shown that the acceptance ratio method and thermodynamic integration are natural generalizations of importance sampling, which is most familiar to statistical audiences.
Unbiased estimation of log normalizing constants with applications to Bayesian cross-validation
• Mathematics, Computer Science
• 2018
This work proposes new unbiased estimators of the logarithm of ratios of normalizing constants of Markov chain Monte Carlo estimators with path sampling, also called thermodynamic integration, and proposes unbiased estimator of the Bayesian cross-validation criterion.
Pareto Smoothed Importance Sampling
• Economics
• 2015
Importance weighting is a general way to adjust Monte Carlo integration to account for draws from the wrong distribution, but the resulting estimate can be noisy when the importance ratios have a
Annealing Markov chain Monte Carlo with applications to ancestral inference
• Mathematics, Computer Science
• 1995
This work proposes MCMC methods distantly related to simulated annealing, which simulate realizations from a sequence of distributions, allowing the distribution being simulated to vary randomly over time.
Pseudo-Extended Markov chain Monte Carlo
• Computer Science
NeurIPS
• 2019
The pseudo-extended MCMC method is introduced as a simple approach for improving the mixing of the MCMC sampler for multi-modal posterior distributions by augments the state-space of the posterior using pseudo-samples as auxiliary variables.