arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2204.01210cs.CV
收藏

协同教学用于无监督域适应与扩展

Co-Teaching for Unsupervised Domain Adaptation and Expansion

Renmin University of China(中国人民大学)

Hailan Lin, Qijie Wei, Kaibin Tian, Ruixiang Zhao, Xirong Li

AI总结:

本文提出协同教学方法,通过知识蒸馏和mixup技术解决跨域视觉模糊性问题,提升无监督域扩展的性能。

摘要:

无监督域适应(UDA)本质上是以牺牲源域上的模型性能来换取目标域上的性能提升。为克服这一问题,引入了无监督域扩展(UDE),该方法在适应目标域的同时保持源域上的性能。在UDA和UDE中,假设针对给定域的模型能够很好地处理该域的样本。我们通过报告跨域视觉模糊性的存在来质疑这一假设:由于两个域之间的边界不明确,一个域的样本可能在视觉上接近另一个域。此类样本通常在其所属域中占少数,因此往往被域特定模型忽视,但可以被另一个域的模型更好地处理。我们利用这一发现,提出了协同教学(CT),其实例化为基于知识蒸馏的CT(kdCT)和基于mixup的CT(miCT)。具体而言,kdCT利用双教师架构来增强学生网络处理跨域模糊性的能力。同时,miCT进一步增强了学生网络的泛化能力。在图像分类和驾驶场景分割的大量实验中展示了CT在UDE中的有效性。

英文摘要:

Unsupervised Domain Adaptation (UDA) essentially trades a model's performance on a source domain for improving its performance on a target domain. To overcome this, Unsupervised Domain Expansion (UDE) has been introduced, which adapts the model to the target domain while preserving its performance in the source domain. In both UDA and UDE, a model tailored to a given domain is assumed to well handle samples from the given domain. We question the assumption by reporting the existence of cross-domain visual ambiguity: Due to the unclear boundary between the two domains, samples from one domain can be visually close to the other domain. Such sorts of samples are typically in the minority in their host domain, so they tend to be overlooked by the domain-specific model, but can be better handled by a model from the other domain. We exploit this finding by proposing Co-Teaching (CT), which is instantiated with knowledge distillation based CT (kdCT) plus mixup based CT (miCT). Specifically, kdCT leverages a dual-teacher architecture to enhance the student network's ability to handle cross-domain ambiguity. Meanwhile, miCT further enhances the generalization ability of the student. Extensive experiments on image classification and driving-scene segmentation show the viability of CT for UDE.

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。

↑