Test-Time Defense Against Adversarial Attacks via Stochastic Resonance of Latent Ensembles
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866908574703157248 |
|---|---|
| author | Lao, Dong Zhang, Yuxiang Oskouie, Haniyeh Ehsani Wu, Yangchao Wong, Alex Soatto, Stefano |
| author_facet | Lao, Dong Zhang, Yuxiang Oskouie, Haniyeh Ehsani Wu, Yangchao Wong, Alex Soatto, Stefano |
| contents | We propose a test-time defense mechanism against adversarial attacks: imperceptible image perturbations that significantly alter the predictions of a model. Unlike existing methods that rely on feature filtering or smoothing, which can lead to information loss, we propose to "combat noise with noise" by leveraging stochastic resonance to enhance robustness while minimizing information loss. Our approach introduces small translational perturbations to the input image, aligns the transformed feature embeddings, and aggregates them before mapping back to the original reference image. This can be expressed in a closed-form formula, which can be deployed on diverse existing network architectures without introducing additional network modules or fine-tuning for specific attack types. The resulting method is entirely training-free, architecture-agnostic, and attack-agnostic. Empirical results show state-of-the-art robustness on image classification and, for the first time, establish a generic test-time defense for dense prediction tasks, including stereo matching and optical flow, highlighting the method's versatility and practicality. Specifically, relative to clean (unperturbed) performance, our method recovers up to 68.1% of the accuracy loss on image classification, 71.9% on stereo matching, and 29.2% on optical flow under various types of adversarial attacks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_03224 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Test-Time Defense Against Adversarial Attacks via Stochastic Resonance of Latent Ensembles Lao, Dong Zhang, Yuxiang Oskouie, Haniyeh Ehsani Wu, Yangchao Wong, Alex Soatto, Stefano Computer Vision and Pattern Recognition Artificial Intelligence Machine Learning We propose a test-time defense mechanism against adversarial attacks: imperceptible image perturbations that significantly alter the predictions of a model. Unlike existing methods that rely on feature filtering or smoothing, which can lead to information loss, we propose to "combat noise with noise" by leveraging stochastic resonance to enhance robustness while minimizing information loss. Our approach introduces small translational perturbations to the input image, aligns the transformed feature embeddings, and aggregates them before mapping back to the original reference image. This can be expressed in a closed-form formula, which can be deployed on diverse existing network architectures without introducing additional network modules or fine-tuning for specific attack types. The resulting method is entirely training-free, architecture-agnostic, and attack-agnostic. Empirical results show state-of-the-art robustness on image classification and, for the first time, establish a generic test-time defense for dense prediction tasks, including stereo matching and optical flow, highlighting the method's versatility and practicality. Specifically, relative to clean (unperturbed) performance, our method recovers up to 68.1% of the accuracy loss on image classification, 71.9% on stereo matching, and 29.2% on optical flow under various types of adversarial attacks. |
| title | Test-Time Defense Against Adversarial Attacks via Stochastic Resonance of Latent Ensembles |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence Machine Learning |
| url | https://arxiv.org/abs/2510.03224 |