Cooperative Distribution Alignment via JSD Upper Bound-猫眼课题宝

DOI：

10.48550/arxiv.2207.02286

发表时间：

2022-07

期刊：

ArXiv

影响因子：

通讯作者：

中科院分区：

其他

文献类型：

作者：

Wonwoong Cho;Ziyu Gong;David I. Inouye

无监督分布对齐估计将两个或更多个源分布映射到共享对齐分布的转换，仅给定每个分布的样本。该任务有许多应用，包括生成建模，无监督域自适应和社会意识学习。大多数先前的作品使用对抗学习（即，最小-最大优化），这对于优化和评估可能是具有挑战性的。一些最近的作品探索了基于非对抗性流（即，可逆）方法，但它们缺乏统一的视角，并且在有效地对齐多个分布方面受到限制。因此，我们建议在一个单一的非对抗性框架下统一和推广以前的基于流的方法，我们证明这相当于最小化Jensen-Shannon Divergence（JSD）的上界。重要的是，我们的问题简化为一个极小，即，合作，问题，并可以提供一个自然的评价指标，无监督的分布对齐。我们在模拟和真实世界的数据集上展示了实证结果，以证明我们的方法的好处。代码可在https://github.com/inouye-lab/alignment-upper-bound上获得。

Unsupervised distribution alignment estimates a transformation that maps two or more source distributions to a shared aligned distribution given only samples from each distribution. This task has many applications including generative modeling, unsupervised domain adaptation, and socially aware learning. Most prior works use adversarial learning (i.e., min-max optimization), which can be challenging to optimize and evaluate. A few recent works explore non-adversarial flow-based (i.e., invertible) approaches, but they lack a unified perspective and are limited in efficiently aligning multiple distributions. Therefore, we propose to unify and generalize previous flow-based approaches under a single non-adversarial framework, which we prove is equivalent to minimizing an upper bound on the Jensen-Shannon Divergence (JSD). Importantly, our problem reduces to a min-min, i.e., cooperative, problem and can provide a natural evaluation metric for unsupervised distribution alignment. We show empirical results on both simulated and real-world datasets to demonstrate the benefits of our approach. Code is available at https://github.com/inouye-lab/alignment-upper-bound.

Cooperative Distribution Alignment via JSD Upper Bound

文献摘要

相似文献