Contradictory Structure Learning for Semi-supervised Domain Adaptation
arxiv(2020)
摘要
Current adversarial adaptation methods attempt to align the cross-domain
features, whereas two challenges remain unsolved: 1) the conditional
distribution mismatch and 2) the bias of the decision boundary towards the
source domain. To solve these challenges, we propose a novel framework for
semi-supervised domain adaptation by unifying the learning of opposite
structures (UODA). UODA consists of a generator and two classifiers (i.e., the
source-scattering classifier and the target-clustering classifier), which are
trained for contradictory purposes. The target-clustering classifier attempts
to cluster the target features to improve intra-class density and enlarge
inter-class divergence. Meanwhile, the source-scattering classifier is designed
to scatter the source features to enhance the decision boundary's smoothness.
Through the alternation of source-feature expansion and target-feature
clustering procedures, the target features are well-enclosed within the dilated
boundary of the corresponding source features. This strategy can make the
cross-domain features to be precisely aligned against the source bias
simultaneously. Moreover, to overcome the model collapse through training, we
progressively update the measurement of feature's distance and their
representation via an adversarial training paradigm. Extensive experiments on
the benchmarks of DomainNet and Office-home datasets demonstrate the
superiority of our approach over the state-of-the-art methods.
更多查看译文
关键词
structure learning,adaptation,semi-supervised
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要