arXiv：cs.LG（机器学习，全量分类）· Andreas Burger, Malte Franke, Luka Mucko, Kjell Jorner, Alan Aspuru-Guzik·· 14 小时前AI 评分31
Grand Canonical Generators：将 Boltzmann 生成器扩展到巨正则系综
Grand Canonical Generators
AI 导读
研究者提出 Grand Canonical Generators（GCG），将 Boltzmann 生成器扩展到巨正则系综，并给出两种设计：一种以化学势为条件联合采样粒子数与构型，另一种将巨正则分布分解为粒子数分布与对应正则 Boltzmann 密度。
正文
Abstract:We introduce Grand Canonical Generators (GCG), a generative framework that extends Boltzmann generators to the grand canonical ensemble. We present two designs. The first conditions a variable-size generative model on the chemical potential, sampling particle number and configuration jointly. The second factorizes the grand canonical distribution into a particle-number distribution and the corresponding canonical Boltzmann density. This factorized formulation can use any existing Boltzmann generator for the canonical component, encodes the known linear chemical-potential dependence analytically, and yields a tractable likelihood that supports self-normalized importance sampling (SNIS). Empirically, GCG accurately reproduces grand canonical observables on a Lennard--Jones fluid and methane adsorption in a zeolite, demonstrating generalization across chemical potentials and correction via SNIS and grand canonical Monte Carlo.
| Comments: | SimBioChem NeurIPS 206 |
| Subjects: | Machine Learning (cs.LG); Machine Learning (stat.ML) |
| Cite as: | arXiv:2610.00683 [cs.LG] |
| (or arXiv:2610.00683v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.00683 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Andreas Burger [view email]
[v1]
Wed, 30 Sep 2026 20:25:15 UTC (6,356 KB)
来源：arXiv：cs.LG（机器学习，全量分类） · arxiv.org