arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2007.04171cs.CVcs.LG
收藏

基于辅助目标领域分类器的领域自适应

Domain Adaptation with Auxiliary Target Domain-Oriented Classifier

National University of Singapore(新加坡国立大学);Sea AI Lab(Sea AI 实验室)

Jian Liang, Dapeng Hu, Jiashi Feng

摘要:

领域自适应(DA)旨在将知识从标签丰富但异质的领域转移到标签稀缺的领域,它简化了标签工作,引起了广泛的关注。不同于以往的方法侧重于学习域不变的特征表示,最近的一些方法提出了通用的半监督学习(SSL)技术,并直接将其应用到DA任务,甚至实现竞争力的性能。最流行的SSL技术之一是伪标签,它通过由标记数据训练的分类器为每个未标记数据分配伪标签。然而,它忽略了DA问题中的分布偏移,不可避免地偏向于源数据。为了解决这个问题,我们提出了一个新的伪标记框架,称为辅助目标领域导向分类器(ATDOC)。ATDOC通过引入仅针对目标数据的辅助分类器来消除分类器的偏差,从而提高伪标签的质量。具体来说,我们采用记忆机制,并开发两种类型的非参数分类器,即最近的质心分类器和邻域聚合,而不引入任何额外的网络参数。尽管其简单的伪分类目标,ATDOC与邻域聚合显着优于域对齐技术和以前的SSL技术在各种各样的DA基准,甚至害怕标记的SSL任务。

英文摘要:

Domain adaptation (DA) aims to transfer knowledge from a label-rich but heterogeneous domain to a label-scare domain, which alleviates the labeling efforts and attracts considerable attention. Different from previous methods focusing on learning domain-invariant feature representations, some recent methods present generic semi-supervised learning (SSL) techniques and directly apply them to DA tasks, even achieving competitive performance. One of the most popular SSL techniques is pseudo-labeling that assigns pseudo labels for each unlabeled data via the classifier trained by labeled data. However, it ignores the distribution shift in DA problems and is inevitably biased to source data. To address this issue, we propose a new pseudo-labeling framework called Auxiliary Target Domain-Oriented Classifier (ATDOC). ATDOC alleviates the classifier bias by introducing an auxiliary classifier for target data only, to improve the quality of pseudo labels. Specifically, we employ the memory mechanism and develop two types of non-parametric classifiers, i.e. the nearest centroid classifier and neighborhood aggregation, without introducing any additional network parameters. Despite its simplicity in a pseudo classification objective, ATDOC with neighborhood aggregation significantly outperforms domain alignment techniques and prior SSL techniques on a large variety of DA benchmarks and even scare-labeled SSL tasks.

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。

↑