Optimization of Annealed Importance Sampling Hyperparameters

TL;DR

This paper introduces a flexible, parameterized Annealed Importance Sampling method that optimizes hyperparameters and intermediary distributions to improve marginal likelihood estimation in deep generative models, especially under limited computational budgets.

Contribution

It proposes a novel parameteric AIS process with residual density-based intermediary distributions, enabling hyperparameter sharing, fixed schedules, and amortized hyperparameter selection.

Findings

Improved marginal likelihood estimates with fewer sampling iterations.

Enhanced performance over traditional AIS methods.

Effective hyperparameter optimization in deep generative models.

Abstract

Annealed Importance Sampling (AIS) is a popular algorithm used to estimates the intractable marginal likelihood of deep generative models. Although AIS is guaranteed to provide unbiased estimate for any set of hyperparameters, the common implementations rely on simple heuristics such as the geometric average bridging distributions between initial and the target distribution which affect the estimation performance when the computation budget is limited. In order to reduce the number of sampling iterations, we present a parameteric AIS process with flexible intermediary distributions defined by a residual density with respect to the geometric mean path. Our method allows parameter sharing between annealing distributions, the use of fix linear schedule for discretization and amortization of hyperparameter selection in latent variable models. We assess the performance of Optimized-Path AIS for marginal likelihood estimation of deep generative models and compare it to compare it to more computationally intensive AIS.