arXiv：cs.LG（机器学习，全量分类）· Jiajun He, Denis Blessing, Mouyang Cheng, Yuanqi Du, Carles Domingo-Enrich·· 14 小时前AI 评分32
Discrete Gibbs Iterative Neural Sampler：离散空间上的定点神经采样器
Fixed-point neural samplers on discrete spaces
AI 导读
研究者提出 Discrete Gibbs Iterative Neural Sampler，一种定点神经采样器，用于从离散、未归一化分布中采样，可显著减少模式坍缩并支持高效可扩展训练。该框架基于 masked diffusion，并可扩展到分布对之间的传输，能有效扩展至高维系统、支持跨不同条件的摊销采样，并实现合金相图的准确估计。
正文
Abstract:Sampling from discrete, unnormalized distributions without access to data is a challenging problem. Neural samplers offer a promising approach by training generative models from density evaluations directly. Despite recent progress, existing discrete neural samplers are prone to mode collapse, come without convergence guarantees when trained via fixed-point iterations, and are often tied to a specific reference process such as masked or uniform diffusion. In this work, we introduce Discrete Gibbs Iterative Neural Sampler, a fixed-point neural sampler that addresses these limitations, enabling efficient, scalable learning, substantially reducing mode collapse in practice. Our framework builds on masked diffusion and also extends to transport between pairs of distributions. We demonstrate that the resulting method scales effectively to high-dimensional systems, supports amortized sampling across different conditions, and enables accurate estimation of alloy phase diagrams.
| Subjects: | Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.01739 [cs.LG] |
| (or arXiv:2610.01739v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.01739 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Denis Blessing [view email]
[v1]
Thu, 1 Oct 2026 14:11:02 UTC (1,055 KB)
来源：arXiv：cs.LG（机器学习，全量分类） · arxiv.org