BriefGPT.xyz
Jun, 2024
Ai-Sampler: 用逆映射进行马尔可夫核的对抗学习
Ai-Sampler: Adversarial Learning of Markov kernels with involutive maps
HTML
PDF
Evgenii Egorov, Ricardo Valperga, Efstratios Gavves
TL;DR
用可逆神经网络构建的逆变 Metropolis-Hastings 核函数以保证细致平衡，通过参数化和训练马尔可夫链的转移核函数来实现高效采样和良好的混合，以最小化链的稳态分布与数据的经验分布之间的总变差距离。
Abstract
markov chain monte carlo methods
have become popular in statistics as versatile techniques to sample from complicated probability distributions. In this work, we propose a method to parameterize and train
transition ker
→