{"id":"arxiv/2610.16248","url":"https://arxiv.org/abs/2610.16248","tags":["cs.LG","cs.AI","math.ST"],"score":0,"title":"Reflected Proximal Dynamics for Accelerated High-Accuracy Log-Concave Sampling","domain":"cs.LG","source":"arxiv","authors":["Bo Yuan","Wei Guo","Yongxin Chen"],"reviews":0,"subject":"cs.LG","summary":"We study sampling from $π(x)\\propto\\mathrm{e}^{-V(x)}$ on $\\mathbb{R}^d$, where $V$ is $α$-strongly convex and $β$-smooth with condition number $κ=β/α$. For this problem, the proximal sampler preserves the target exactly and recovers overdamped Langevin dynamics in the small-step limit, but its Gaussian resampling step discards directional information. Underdamped Langevin dynamics retains such information through momentum to accelerate exploration, motivating a proximal construction that combines target preservation with directional persistence. We introduce \\textit{reflected proximal…","category":"research","comments":0,"disputed":false,"division":"cs","mentions":0,"discussed":0,"published":"2026-10-12","reactions":{}}