Reducing the representational discrepancy between source and target domains is a key component to maximize the model generalization. In this work, we advocate for leveraging natural language supervision for the domain generalization task. We introduce two modules to ground visual representations with texts containing typical reasoning of humans: (1) Visual and Textual Joint Embedder and (2) Textual Explanation Generator. The former learns the image-text joint embedding space where we can ground high-level class-discriminative information into the model. The latter leverages an explainable model and generates explanations justifying the rationale behind its decision. To the best of our knowledge, this is the first work to leverage the vision-and-language cross-modality approach for the domain generalization task. Our experiments with a newly created CUB-DG benchmark dataset demonstrate that cross-modality supervision can be successfully used to ground domain-invariant visual representations and improve the model generalization. Furthermore, in the large-scale DomainBed benchmark, our proposed method achieves state-of-the-art results and ranks 1st in average performance for five multi-domain datasets. The dataset and codes are available at https://github.com/mswzeus/GVRT.
|Title of host publication||Computer Vision – ECCV 2022 - 17th European Conference, Proceedings|
|Editors||Shai Avidan, Gabriel Brostow, Moustapha Cissé, Giovanni Maria Farinella, Tal Hassner|
|Publisher||Springer Science and Business Media Deutschland GmbH|
|Number of pages||17|
|Publication status||Published - 2022|
|Event||17th European Conference on Computer Vision, ECCV 2022 - Tel Aviv, Israel|
Duration: 2022 Oct 23 → 2022 Oct 27
|Name||Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)|
|Conference||17th European Conference on Computer Vision, ECCV 2022|
|Period||22/10/23 → 22/10/27|
Bibliographical noteFunding Information:
Acknowledgements. This work was supported by supported by the National Research Foundation of Korea grant (NRF-2021R1C1C1009608), Basic Science Research Program (NRF-2021R1A6A1A13044830), and ICT Creative Consilience program (IITP-2022-2022-0-01819).
© 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.
- Domain generalization
- Image classification
- Textual explanation
- Visual-textual joint embedding
ASJC Scopus subject areas
- Theoretical Computer Science
- Computer Science(all)