Branching Stein Variational Gradient Descent for sampling multimodal distributions
This addresses a challenge in variational inference for researchers and practitioners dealing with complex, multimodal data, though it appears incremental as an extension of SVGD.
The paper tackles the problem of sampling from multimodal distributions by proposing Branched Stein Variational Gradient Descent (BSVGD), which extends SVGD with a random branching mechanism to improve exploration, resulting in validated performance through numerical experiments and theoretical convergence guarantees.
We propose a novel particle-based variational inference method designed to work with multimodal distributions. Our approach, referred to as Branched Stein Variational Gradient Descent (BSVGD), extends the classical Stein Variational Gradient Descent (SVGD) algorithm by incorporating a random branching mechanism that encourages the exploration of the state space. In this work, a theoretical guarantee for the convergence in distribution is presented, as well as numerical experiments to validate the suitability of our algorithm. Performance comparisons between the BSVGD and the SVGD are presented using the Wasserstein distance between samples and the corresponding computational times.