Elias Hess-Childs, Dejan Slepčev, Lantian Xu
The paper introduces new Radon--Wasserstein gradient flows for efficient high-dimensional sampling using interacting particles with linear scaling costs.
The study presents a new method for sampling from complex probability distributions, which is important in fields like statistics and machine learning. The method uses a mathematical concept called gradient flows to gradually transform a simple distribution into a target distribution. This approach is efficient even in high-dimensional spaces, thanks to a novel use of geometry and efficient computational techniques. The authors show that their method works well in practice and provide mathematical proofs to support its effectiveness.
