arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2207.04564cs.CLcs.LG
收藏

无监督领域自适应的领域混淆对比学习

Domain Confused Contrastive Learning for Unsupervised Domain Adaptation

Nanyang Technological University(南洋理工大学)

Quanyu Long, Tianze Luo, Wenya Wang, Sinno Jialin Pan

摘要:

在这项工作中,我们研究了无监督领域自适应(UDA)在一个具有挑战性的自我监督的方法。其中一个难点是如何在没有目标标签的情况下学习任务区分。与以往文献中直接对齐跨域分布或利用反向梯度不同,我们提出了域混淆对比学习(DCCL),通过域拼图来桥接源域和目标域,并在适应后保留区分性表示。从技术上讲,DCCL搜索最具领域挑战性的方向,并将领域混淆的增强作为正对进行精细加工,然后对比鼓励模型将表示拉向另一个领域,从而学习更稳定和有效的领域不变性。我们还研究了在进行其他数据增强时,对比学习是否一定有助于UDA。大量的实验表明,DCCL显着优于基线。

英文摘要:

In this work, we study Unsupervised Domain Adaptation (UDA) in a challenging self-supervised approach. One of the difficulties is how to learn task discrimination in the absence of target labels. Unlike previous literature which directly aligns cross-domain distributions or leverages reverse gradient, we propose Domain Confused Contrastive Learning (DCCL) to bridge the source and the target domains via domain puzzles, and retain discriminative representations after adaptation. Technically, DCCL searches for a most domain-challenging direction and exquisitely crafts domain confused augmentations as positive pairs, then it contrastively encourages the model to pull representations towards the other domain, thus learning more stable and effective domain invariances. We also investigate whether contrastive learning necessarily helps with UDA when performing other data augmentations. Extensive experiments demonstrate that DCCL significantly outperforms baselines.

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。

↑