ART

Adversarial Robustness Toolbox (ART) is a Python library for machine learning security. It lets developers and researchers evaluate, defend, certify, and verify machine learning models and applications against adversarial threats such as evasion, poisoning, extraction, and inference. It supports major frameworks such as TensorFlow, Keras, PyTorch, scikit-learn, and XGBoost, all data types including images, tables, audio, and video, and tasks such as classification, object detection, generation, and certification. It is hosted by the LF AI & Data Foundation.

• Collection of evasion attacks (PGD, Carlini-Wagner, DeepFool, etc.)
• Extraction, inference, and poisoning (backdoor) attacks
• Detector, preprocessing, and postprocessing defenses
• Adversarial Training trainer
• Certification estimators (Randomized Smoothing, Interval, etc.)
• Support for frameworks such as TensorFlow, Keras, PyTorch, scikit-learn, and XGBoost
• Support for image, tabular, audio, and video data

Open source + paid enterprise plans

Some open-source tools are free; enterprise security platforms are billed under separate contracts

Estimate based on AI knowledge - please verify on the official site

• One of 21 AI tools in the AI Security & Safety field.
• Connected to 6 other tools — #1 in its field.