search

Trusted-AI / adversarial-robustness-toolbox

Adversarial Robustness Toolbox (ART) - Python Library for Machine Learning Security - Evasion, Poisoning, Extraction, Inference - Red and Blue Teams
https://adversarial-robustness-toolbox.readthedocs.io/en/latest/
MIT License
4.76k stars 1.15k forks source link

Implement `get_activations()` for PyTorch Object Detectors #2314

Open f4str opened 11 months ago

f4str commented 11 months ago

Is your feature request related to a problem? Please describe. The current PyTorch object detectors (PyTorchObjectDetector, PyTorchFasterRCNN, PyTorchYolo, PyTorchDetectionTransformer) do not implement the get_activations() method and therefore cannot extract activations.

Describe the solution you'd like Using the model wrapper method used by HuggingFaceClassifierPyTorch, support can be added for activations for the PyTorch object detectors. This will allow usage of activation-based defenses such as Activation Clustering and Spectral Signatures.

Describe alternatives you've considered N/A

Additional context This will be worked on after #2267 is completed to avoid redundant code.

beat-buesser commented 10 months ago

Hi @f4str Sounds good!