Robust Local Preserving and Global Aligning Network for Adversarial Domain Adaptation

Unsupervised domain adaptation (UDA) requires source samples with clean ground truth labels during training. Accurately labeling a large number of is time-consuming and laborious. An alternative to utilize noisy for However, training can greatly reduce the performance UDA. In this paper, we address problem that learning UDA models only access propose novel method called robust local preserving global aligning network (RLPGA). RLPGA improves robustness label noise from two aspects. One classifier by informative-theoretic-based loss function. The other constructing adjacency weight matrices negative proposed module preserve topology structures input data. We conduct theoretical analysis on prove are beneficial empirical risk target domain. A series studies show effectiveness our RLPGA.