Id: arxiv/2602.17087 Url: https://arxiv.org/abs/2602.17087 Tags: math.PR, stat.CO Score: 0 Title: Diffusive Scaling Limits of Forward Event-Chain Monte Carlo: Provably Efficient Exploration with Partial Refreshment Domain: math.PR Source: arxiv Authors: Hirofumi Shiba, Kengo Kamatani Reviews: 0 Subject: math.PR Category: research Comments: 0 Disputed: false Division: math Mentions: 0 Discussed: 0 Published: 2026-02-19 Page: https://delphi.science/r/arxiv/2602.17087 Piecewise deterministic Markov process samplers are attractive alternatives to Metropolis--Hastings algorithms. A central design question is how to incorporate partial velocity refreshment to ensure ergodicity without injecting excessive noise. Forward Event-Chain Monte Carlo (FECMC) is a generalization of the Bouncy Particle Sampler (BPS) that addresses this issue through a stochastic reflection mechanism, thereby reducing reliance on global refreshment moves. Despite promising empirical performance, its theoretical efficiency remains largely unexplored. We develop a high-dimensional…