Prompting the Unseen: Detecting Hidden Backdoors in Black-Box Models
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2024
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866912312478138368 |
|---|---|
| author | Huang, Zi-Xuan Chen, Jia-Wei Zhang, Zhi-Peng Yu, Chia-Mu |
| author_facet | Huang, Zi-Xuan Chen, Jia-Wei Zhang, Zhi-Peng Yu, Chia-Mu |
| contents | Visual prompting (VP) is a new technique that adapts well-trained frozen models for source domain tasks to target domain tasks. This study examines VP's benefits for black-box model-level backdoor detection. The visual prompt in VP maps class subspaces between source and target domains. We identify a misalignment, termed class subspace inconsistency, between clean and poisoned datasets. Based on this, we introduce \textsc{BProm}, a black-box model-level detection method to identify backdoors in suspicious models, if any. \textsc{BProm} leverages the low classification accuracy of prompted models when backdoors are present. Extensive experiments confirm \textsc{BProm}'s effectiveness. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2411_09540 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Prompting the Unseen: Detecting Hidden Backdoors in Black-Box Models Huang, Zi-Xuan Chen, Jia-Wei Zhang, Zhi-Peng Yu, Chia-Mu Computer Vision and Pattern Recognition Artificial Intelligence Machine Learning Visual prompting (VP) is a new technique that adapts well-trained frozen models for source domain tasks to target domain tasks. This study examines VP's benefits for black-box model-level backdoor detection. The visual prompt in VP maps class subspaces between source and target domains. We identify a misalignment, termed class subspace inconsistency, between clean and poisoned datasets. Based on this, we introduce \textsc{BProm}, a black-box model-level detection method to identify backdoors in suspicious models, if any. \textsc{BProm} leverages the low classification accuracy of prompted models when backdoors are present. Extensive experiments confirm \textsc{BProm}'s effectiveness. |
| title | Prompting the Unseen: Detecting Hidden Backdoors in Black-Box Models |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence Machine Learning |
| url | https://arxiv.org/abs/2411.09540 |