Weakly Supervised Contrastive Adversarial Training for Learning Robust Features from Semi-supervised Data
Fuente:
arXiv
Guardado en:
| Autores principales: | , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2025
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866909544755494912 |
|---|---|
| author | Zhang, Lilin Wu, Chengpei Yang, Ning |
| author_facet | Zhang, Lilin Wu, Chengpei Yang, Ning |
| contents | Existing adversarial training (AT) methods often suffer from incomplete perturbation, meaning that not all non-robust features are perturbed when generating adversarial examples (AEs). This results in residual correlations between non-robust features and labels, leading to suboptimal learning of robust features. However, achieving complete perturbation, i.e., perturbing as many non-robust features as possible, is challenging due to the difficulty in distinguishing robust and non-robust features and the sparsity of labeled data. To address these challenges, we propose a novel approach called Weakly Supervised Contrastive Adversarial Training (WSCAT). WSCAT ensures complete perturbation for improved learning of robust features by disrupting correlations between non-robust features and labels through complete AE generation over partially labeled data, grounded in information theory. Extensive theoretical analysis and comprehensive experiments on widely adopted benchmarks validate the superiority of WSCAT. Our code is available at https://github.com/zhang-lilin/WSCAT. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2503_11032 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Weakly Supervised Contrastive Adversarial Training for Learning Robust Features from Semi-supervised Data Zhang, Lilin Wu, Chengpei Yang, Ning Computer Vision and Pattern Recognition Existing adversarial training (AT) methods often suffer from incomplete perturbation, meaning that not all non-robust features are perturbed when generating adversarial examples (AEs). This results in residual correlations between non-robust features and labels, leading to suboptimal learning of robust features. However, achieving complete perturbation, i.e., perturbing as many non-robust features as possible, is challenging due to the difficulty in distinguishing robust and non-robust features and the sparsity of labeled data. To address these challenges, we propose a novel approach called Weakly Supervised Contrastive Adversarial Training (WSCAT). WSCAT ensures complete perturbation for improved learning of robust features by disrupting correlations between non-robust features and labels through complete AE generation over partially labeled data, grounded in information theory. Extensive theoretical analysis and comprehensive experiments on widely adopted benchmarks validate the superiority of WSCAT. Our code is available at https://github.com/zhang-lilin/WSCAT. |
| title | Weakly Supervised Contrastive Adversarial Training for Learning Robust Features from Semi-supervised Data |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2503.11032 |