Domain-Specific Bias Filtering for Single Labeled Domain Generalization

10/02/2021
by   Junkun Yuan, et al.
0

Domain generalization (DG) utilizes multiple labeled source datasets to train a generalizable model for unseen target domains. However, due to expensive annotation costs, the requirements of labeling all the source data are hard to be met in real-world applications. In this paper, we investigate a Single Labeled Domain Generalization (SLDG) task with only one source domain being labeled, which is more practical and challenging than the Conventional Domain Generalization (CDG). A major obstacle in the SLDG task is the discriminability-generalization bias: discriminative information in the labeled source dataset may contain domain-specific bias, constraining the generalization of the trained model. To tackle this challenging task, we propose a novel method called Domain-Specific Bias Filtering (DSBF), which initializes a discriminative model with the labeled source data and filters out its domain-specific bias with the unlabeled source data for generalization improvement. We divide the filtering process into: (1) Feature extractor debiasing using k-means clustering-based semantic feature re-extraction; and (2) Classifier calibrating using attention-guided semantic feature projection. DSBF unifies the exploration of the labeled and the unlabeled source data to enhance the discriminability and generalization of the trained model, resulting in a highly generalizable model. We further provide theoretical analysis to verify the proposed domain-specific bias filtering process. Extensive experiments on multiple datasets show the superior performance of DSBF in tackling both the challenging SLDG task and the CDG task.

READ FULL TEXT

Please sign up or login with your details

Forgot password? Click here to reset