{"id":"arxiv/2603.09251","url":"https://arxiv.org/abs/2603.09251","tags":["stat.ML","cs.LG","math.NA"],"score":0,"title":"A Generative Sampler for distributions with possible discrete parameter based on Reversibility","domain":"stat.ML","source":"arxiv","authors":["Lei Li","Zhen Wang","Lishuo Zhang"],"reviews":0,"subject":"stat.ML","summary":"Learning to sample from complex unnormalized distributions is a fundamental challenge in computational physics and machine learning. While score-based and variational methods have achieved success in continuous domains, extending them to discrete or mixed-variable systems remains difficult due to ill-defined gradients or high variance in estimators. We propose a unified, target-gradient-free generative sampling framework applicable across diverse state spaces. Building on the fact that detailed balance implies the time-reversibility of the equilibrium stochastic process, we enforce this…","category":"research","comments":0,"disputed":false,"division":"stat","mentions":0,"discussed":0,"published":"2026-03-10","reactions":{}}