adversarial-robustness-toolbox
adversarial-robustness-toolbox copied to clipboard
Trusted-AI
Adversarial Robustness Toolbox (ART) - Python Library for Machine Learning Security - Evasion, Poisoning, Extraction, Inference - Red and Blue Teams
**Algorithm described here** https://arxiv.org/pdf/1907.05587.pdf **Implementation** As a part of engineering thesis I'm implementing this algorithm. In addition, I will provide example usages of new module including: jupyter notebook and kubeflow...
**Describe the bug** A clear and concise description of what the bug is. **To Reproduce** Steps to reproduce the behavior: 1. Go to 'adversarial_action_recognition.ipynb' 2. Click on 'adv_sample = fgm.generate(...
**Is your feature request related to a problem? Please describe.** Currently attribute inference attacks only support categorical features. **Describe the solution you'd like** At least in the black box and...
**Is your feature request related to a problem? Please describe.** add the option to use an L-infinity bound with the `SimBA` attack **Describe the solution you'd like** adapt for `art`...
### Discussed in https://github.com/Trusted-AI/adversarial-robustness-toolbox/discussions/1383 Originally posted by **dominic-simon** October 29, 2021 As the title states, I'm having issues with the Adversarial Patch Notebook, specifically the TensorFlowV2 notebook. The patch shown...
**Is your feature request related to a problem? Please describe.** I have a tensorflow model which utilizes multiple inputs which are of varying shape. Specifically I have a video input...
**Describe the bug** for image shape as [128, 188, 1], square attack raise value error **To Reproduce** Steps to reproduce the behavior: 1. train a model with images shape as...
We would like to implement support for Expectation over Transformation (EoT) of rotation (in steps of 90 degrees) in object detection applications in PyTorch and maybe TensorFlow.
**Describe the bug** An import in the _get_layers function in the TensorFlowClassifier class overwrites the import when trying to use the tf1 compatibility layer in tf2 **To Reproduce** Steps to...
**Describe the bug** There is an error in cell 5 in [notebooks/detection_adversarial_samples_cifar10.ipynb](https://github.com/Trusted-AI/adversarial-robustness-toolbox/blob/main/notebooks/detection_adversarial_samples_cifar10.ipynb) https://github.com/Trusted-AI/adversarial-robustness-toolbox/blame/main/notebooks/detection_adversarial_samples_cifar10.ipynb#L405-L407 **In [5]:** ```Python x_test_pred = np.argmax(classifier.predict(x_test[:100]), axis=1) nb_correct_pred = np.sum(x_test_pred == np.argmax(y_test[:100], axis=1)) print("Original test data (first...
Metadata
Owner
Metadata
Adversarial Robustness Toolbox (ART) - Python Library for Machine Learning Security - Evasion, Poisoning, Extraction, Inference - Red and Blue Teams