q-Paths: Generalizing the Geometric Annealing Path using Power Means

Abstract

Many common machine learning methods involve the geometric annealing path, a sequence of intermediate densities between two distributions of interest constructed using the geometric average. While alternatives such as the moment-averaging path have demonstrated performance gains in some settings, their practical applicability remains limited by exponential family endpoint assumptions and a lack of closed form energy function. In this work, we introduce $q$-paths, a family of paths which is derived from a generalized notion of the mean, includes the geometric and arithmetic mixtures as special cases, and admits a simple closed form involving the deformed logarithm function from nonextensive thermodynamics. Following previous analysis of the geometric path, we interpret our $q$-paths as corresponding to a $q$-exponential family of distributions, and provide a variational representation of intermediate densities as minimizing a mixture of $\alpha$-divergences to the endpoints. We show that small deviations away from the geometric path yield empirical gains for Bayesian inference using Sequential Monte Carlo and generative model evaluation using Annealed Importance Sampling.

arXiv commentarXiv admin note: text overlap with arXiv:2012.07823

References in corpus (9)

• MCMC using Hamiltonian dynamics
• On a generalization of the Jensen-Shannon divergence and the JS-symmetrization of distances relying on abstract means
• Pareto Smoothed Importance Sampling
• Yes, but Did It Work?: Evaluating Variational Inference
• The q-exponential family in statistical physics
• On the Quantitative Analysis of Decoder-Based Generative Models
• Sandwiching the marginal likelihood using bidirectional Monte Carlo
• Advantages of $q$-logarithm representation over $q$-exponential representation from the sense of scale and shift on nonlinear systems
• Likelihood Ratio Exponential Families