A Broad Study of Pre-training for Domain Generalization and Adaptation
This work addresses the problem of domain transfer for machine learning practitioners by providing insights into pre-training effects, though it is incremental as it builds on existing methods.
The study investigated the impact of pre-training on domain adaptation and generalization, finding that using a state-of-the-art backbone outperformed existing baselines by 10.7% on Office-Home and 5.5% on DomainNet.
Deep models must learn robust and transferable representations in order to perform well on new domains. While domain transfer methods (e.g., domain adaptation, domain generalization) have been proposed to learn transferable representations across domains, they are typically applied to ResNet backbones pre-trained on ImageNet. Thus, existing works pay little attention to the effects of pre-training on domain transfer tasks. In this paper, we provide a broad study and in-depth analysis of pre-training for domain adaptation and generalization, namely: network architectures, size, pre-training loss, and datasets. We observe that simply using a state-of-the-art backbone outperforms existing state-of-the-art domain adaptation baselines and set new baselines on Office-Home and DomainNet improving by 10.7\% and 5.5\%. We hope that this work can provide more insights for future domain transfer research.