{"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","summary":"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…","category":"research","comments":0,"disputed":false,"division":"math","mentions":0,"discussed":0,"published":"2026-02-19","reactions":{}}