All videos
All videos
ART 360: Defending AI models against adversarial attacks
October 8, 2019
Adversarial samples are inputs to Deep Neural Networks (DNNs) that an adversary has tampered with in order to cause misclassifications. It is surprisingly easy to create adversarial samples and surprisingly difficult to defend DNNs against them. In this talk, I will review the state-of-the-art and recent progress in better understanding adversarial samples and developing DNNs that are robust against them. I will then give a perspective on the potential threats that adversarial samples pose to security-critical applications of DNNs. Finally, I will show how researchers and developers can experiment with adversarial attacks and defences using the ART 360 open-source library https://github.com/IBM/adversarial-robustness-toolbox.
Other videos that you might like
How likes influence your views
Paweł Rzeszuciński
Trained scikit-learn model refinery with open-source packages
Dorota Lączak, Maria Oleszkiewicz
Explaining neural networks predictions
Matthew Opala
How to build an AI from zero to learn to play and solve a tough game
Juan Tomás García