Abstract
Sampling from high-dimensional, multi-modal distributions remains afundamental challenge across domains such as statistical Bayesian inference andphysics-based machine learning. In this paper, we propose Annealing Flow (AF),a continuous normalizing flow-based approach designed to sample fromhigh-dimensional and multi-modal distributions. The key idea is to learn acontinuous normalizing flow-based transport map, guided by annealing, totransition samples from an easy-to-sample distribution to the targetdistribution, facilitating effective exploration of modes in high-dimensionalspaces. Unlike many existing methods, AF training does not rely on samples fromthe target distribution. AF ensures effective and balanced mode exploration,achieves linear complexity in sample size and dimensions, and circumventsinefficient mixing times. We demonstrate the superior performance of AFcompared to state-of-the-art methods through extensive experiments on variouschallenging distributions and real-world datasets, particularly inhigh-dimensional and multi-modal settings. We also highlight the potential ofAF for sampling the least favorable distributions.