Adversarial Teacher-Student Representation Learning for Domain Generalization
Journal
Advances in Neural Information Processing Systems
Journal Volume
23
Pages
19448-19460
Date Issued
2021
Author(s)
Abstract
Domain generalization (DG) aims to transfer the learning task from a single or multiple source domains to unseen target domains. To extract and leverage the information which exhibits sufficient generalization ability, we propose a simple yet effective approach of Adversarial Teacher-Student Representation Learning, with the goal of deriving the domain generalizable representations via generating and exploring out-of-source data distributions. Our proposed framework advances Teacher-Student learning in an adversarial learning manner, which alternates between knowledge-distillation based representation learning and novel-domain data augmentation. The former progressively updates the teacher network for deriving domain-generalizable representations, while the latter synthesizes data out-ofsource yet plausible distributions. Extensive image classification experiments on benchmark datasets in multiple and single source DG settings confirm that, our model exhibits sufficient generalization ability and performs favorably against state-of-the-art DG methods. © 2021 Neural information processing systems foundation. All rights reserved.
Other Subjects
Classification (of information); Distillation; Effective approaches; Generalisation; Generalization ability; Learning tasks; Multiple source; Simple++; Single source; Source data; Target domain; Teachers'; Students
Type
conference paper