arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2406.07423cs.LGcs.AIstat.ML
收藏

超越ELBO:变分采样方法的大规模评估

Beyond ELBOs: A Large-Scale Evaluation of Variational Methods for Sampling

Karlsruhe Institute of Technology(卡尔斯鲁厄理工学院);University of Cambridge(剑桥大学);FZI Research Center for Information Technology(FZI信息技术研究中心)

Denis Blessing, Xiaogang Jia, Johannes Esslinger, Francisco Vargas, Gerhard Neumann

AI总结:

本文提出一个标准化基准,在大规模任务上系统评估变分采样方法,并引入新模式坍塌度量,为采样方法的发展提供参考。

摘要:

蒙特卡洛方法、变分推断及其组合在从难以处理的概率分布中采样方面发挥着关键作用。然而,当前研究缺乏统一的评估框架,依赖不同的性能度量,且在不同任务上的方法比较有限,这使得评估进展变得复杂,并妨碍了实践者的决策。针对这些挑战,我们的工作引入了一个基准,该基准使用标准化的任务套件和广泛的性能标准来评估采样方法。此外,我们研究了用于量化模式坍塌的现有度量,并为此引入了新的度量。我们的发现提供了对现有采样方法优势与不足的深入认识,为未来发展提供了有价值的参考。代码已在此公开。

英文摘要:

Monte Carlo methods, Variational Inference, and their combinations play a pivotal role in sampling from intractable probability distributions. However, current studies lack a unified evaluation framework, relying on disparate performance measures and limited method comparisons across diverse tasks, complicating the assessment of progress and hindering the decision-making of practitioners. In response to these challenges, our work introduces a benchmark that evaluates sampling methods using a standardized task suite and a broad range of performance criteria. Moreover, we study existing metrics for quantifying mode collapse and introduce novel metrics for this purpose. Our findings provide insights into strengths and weaknesses of existing sampling methods, serving as a valuable reference for future developments. The code is publicly available here.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑