Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2024
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2407.05765 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866929684876361728 |
|---|---|
| author | Zhu, Yaoyao Cai, Xiuding Wang, Yingkai Miao, Dong Fu, Zhongliang Luo, Xu |
| author_facet | Zhu, Yaoyao Cai, Xiuding Wang, Yingkai Miao, Dong Fu, Zhongliang Luo, Xu |
| contents | Deep learning models excel in computer vision tasks but often fail to generalize to out-of-distribution (OOD) domains. Invariant Risk Minimization (IRM) aims to address OOD generalization by learning domain-invariant features. However, IRM struggles with datasets exhibiting significant diversity shifts. While data augmentation methods like Mixup and Semantic Data Augmentation (SDA) enhance diversity, they risk over-augmentation and label instability. To address these challenges, we propose a domain-shared Semantic Data Augmentation (SDA) module, a novel implementation of Variance Risk Minimization (VRM) designed to enhance dataset diversity while maintaining label consistency. We further provide a Rademacher complexity analysis, establishing a tighter generalization error bound compared to baseline methods. Extensive evaluations on OOD benchmarks, including PACS, VLCS, OfficeHome, and TerraIncognita, demonstrate consistent performance improvements over state-of-the-art domain generalization methods. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2407_05765 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Invariance Principle Meets Vicinal Risk Minimization Zhu, Yaoyao Cai, Xiuding Wang, Yingkai Miao, Dong Fu, Zhongliang Luo, Xu Computer Vision and Pattern Recognition Deep learning models excel in computer vision tasks but often fail to generalize to out-of-distribution (OOD) domains. Invariant Risk Minimization (IRM) aims to address OOD generalization by learning domain-invariant features. However, IRM struggles with datasets exhibiting significant diversity shifts. While data augmentation methods like Mixup and Semantic Data Augmentation (SDA) enhance diversity, they risk over-augmentation and label instability. To address these challenges, we propose a domain-shared Semantic Data Augmentation (SDA) module, a novel implementation of Variance Risk Minimization (VRM) designed to enhance dataset diversity while maintaining label consistency. We further provide a Rademacher complexity analysis, establishing a tighter generalization error bound compared to baseline methods. Extensive evaluations on OOD benchmarks, including PACS, VLCS, OfficeHome, and TerraIncognita, demonstrate consistent performance improvements over state-of-the-art domain generalization methods. |
| title | Invariance Principle Meets Vicinal Risk Minimization |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2407.05765 |