Domain Generalization by Mutual-Information Regularization with Pre-Trained Models

Abstract

Domain generalization (DG) aims to learn a generalized model to an unseen target domain using only limited source domains. Previous attempts to DG fail to learn domain-invariant representations only from the source domains due to the significant domain shifts between training and test domains. Instead, we re-formulate the DG objective using mutual information with the oracle model, a model generalized to any possible domain. We derive a tractable variational lower bound via approximating the oracle model by a pre-trained model, called Mutual Information Regularization with Oracle (MIRO). Our extensive experiments show that MIRO significantly improves the out-of-distribution performance. Furthermore, our scaling experiments show that the larger the scale of the pre-trained model, the greater the performance improvement of MIRO. Code is available at https://github.com/kakaobrain/miro.

Cite

Text

Cha et al. "Domain Generalization by Mutual-Information Regularization with Pre-Trained Models." Proceedings of the European Conference on Computer Vision (ECCV), 2022. doi:10.1007/978-3-031-20050-2_26

Markdown

[Cha et al. "Domain Generalization by Mutual-Information Regularization with Pre-Trained Models." Proceedings of the European Conference on Computer Vision (ECCV), 2022.](https://mlanthology.org/eccv/2022/cha2022eccv-domain/) doi:10.1007/978-3-031-20050-2_26

BibTeX

@inproceedings{cha2022eccv-domain,
  title     = {{Domain Generalization by Mutual-Information Regularization with Pre-Trained Models}},
  author    = {Cha, Junbum and Lee, Kyungjae and Park, Sungrae and Chun, Sanghyuk},
  booktitle = {Proceedings of the European Conference on Computer Vision (ECCV)},
  year      = {2022},
  doi       = {10.1007/978-3-031-20050-2_26},
  url       = {https://mlanthology.org/eccv/2022/cha2022eccv-domain/}
}