Non-Robust Features are Not Always Useful in One-Class Classification
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2024
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866929414383599616 |
|---|---|
| author | Lau, Matthew Wang, Haoran Helbling, Alec Hul, Matthew Peng, ShengYun Andreoni, Martin Lunardi, Willian T. Lee, Wenke |
| author_facet | Lau, Matthew Wang, Haoran Helbling, Alec Hul, Matthew Peng, ShengYun Andreoni, Martin Lunardi, Willian T. Lee, Wenke |
| contents | The robustness of machine learning models has been questioned by the existence of adversarial examples. We examine the threat of adversarial examples in practical applications that require lightweight models for one-class classification. Building on Ilyas et al. (2019), we investigate the vulnerability of lightweight one-class classifiers to adversarial attacks and possible reasons for it. Our results show that lightweight one-class classifiers learn features that are not robust (e.g. texture) under stronger attacks. However, unlike in multi-class classification (Ilyas et al., 2019), these non-robust features are not always useful for the one-class task, suggesting that learning these unpredictive and non-robust features is an unwanted consequence of training. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2407_06372 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Non-Robust Features are Not Always Useful in One-Class Classification Lau, Matthew Wang, Haoran Helbling, Alec Hul, Matthew Peng, ShengYun Andreoni, Martin Lunardi, Willian T. Lee, Wenke Machine Learning Computer Vision and Pattern Recognition 68T45 I.2.10; I.4.10; I.5.4 The robustness of machine learning models has been questioned by the existence of adversarial examples. We examine the threat of adversarial examples in practical applications that require lightweight models for one-class classification. Building on Ilyas et al. (2019), we investigate the vulnerability of lightweight one-class classifiers to adversarial attacks and possible reasons for it. Our results show that lightweight one-class classifiers learn features that are not robust (e.g. texture) under stronger attacks. However, unlike in multi-class classification (Ilyas et al., 2019), these non-robust features are not always useful for the one-class task, suggesting that learning these unpredictive and non-robust features is an unwanted consequence of training. |
| title | Non-Robust Features are Not Always Useful in One-Class Classification |
| topic | Machine Learning Computer Vision and Pattern Recognition 68T45 I.2.10; I.4.10; I.5.4 |
| url | https://arxiv.org/abs/2407.06372 |