发表机构
MBZUAI; School of Applied Mathematics, Getulio Vargas Foundation（穆罕默德·本·扎耶德人工智能大学; 热图利奥·瓦加斯基金会应用数学学院）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
本文证明生成式边缘化模型与GFlowNets等价，并提出基于Gelman-Rubin统计量的自动再生准则，扩展采样至非自回归过程，所提粒子GFlowNets在大型组合空间中显著加速训练。
AI 中文摘要
生成式边缘化模型（MaMs）最近被引入为用于离散分布任意阶自回归建模的高效神经采样模型。通过学习持久块吉布斯采样器的边缘概率和条件概率，MaMs能够通过单次神经网络前向传播实现快速后验评估。虽然先前的工作认为MaMs与生成流网络（GFlowNets）——一种在离散随机模型中推理的成熟范式——是不同的，但我们证明它们是等价的。然后，我们还将MaMs的采样策略扩展到非自回归生成过程。特别地，我们描述了一种基于Gelman-Rubin统计量的全状态再生的自动准则，该准则在加速学习收敛中起关键作用。我们的实验表明，我们提出的方法，称为粒子GFlowNets，在大型组合空间中显著加速了训练。
英文摘要
Generative Marginalization Models (MaMs) have been recently introduced as efficient neural sampling models for any-order autoregressive modelling of discrete distributions. By learning both the marginal and conditional probabilities of a persistent-block Gibbs sampler, MaMs enable fast posterior evaluation with a single neural network forward pass. While prior work has considered MaMs to be distinct from Generative Flow Networks (GFlowNets), a well-established paradigm for inference in discrete stochastic models, we show that they are equivalent. Then, we also extend MaMs' sampling strategy to non-autoregressive generative processes. In particular, we describe an automatic criterion for full-state rejuvenation of the Gibbs sampler, derived from the Gelman-Rubin statistic, which plays a key role in speeding up learning convergence. Our experiments show that our method, called Particle GFlowNets, markedly accelerates training in large combinatorial spaces.
CommentsAccepted at UAI 2026
