Improved off-policy training of diffusion samplers
Mila, Université de Montréal, Jagiellonian University · KAIST / Mila - Quebec AI Institute · Universite de Montreal / MILA · Montreal Institute for Learning Algorithms, University of Montreal, Université de Montréal · Mila AI Institute · Montreal Institute for Learning Algorithms, University of Montreal, University of Montreal · Université de Montréal · Mila/U. Montreal · University of Edinburgh
PDF 由论文原始站点提供，PaperCompass 不保存论文文件。
摘要
We study the problem of training diffusion models to sample from a distribution with a given unnormalized density or energy function. We benchmark several diffusion-structured inference methods, including simulation-based variational approaches and off-policy methods (continuous generative flow networks). Our results shed light on the relative advantages of existing algorithms while bringing into question some claims from past work. We also propose a novel exploration strategy for off-policy methods, based on local search in the target space with the use of a replay buffer, and show that it improves the quality of samples on a variety of target distributions. Our code for the sampling methods and benchmarks studied is made public at [this link](https://github.com/GFNOrg/gfn-diffusion) as a base for future work on diffusion models for amortized inference.