Session

With the advance in Artificial Intelligence, neural networks are everywhere! But like all technologies, it has its flaws. Adversarial examples are able to fool neural networks used for critical applications such as self-driving cars, face recognition apps, security cameras… Together, we will see how to generate your own adversarial images and fool a neural network into seeing things that are not there. We'll also see how to defend against adversarial attacks. No Tensorflow, there is no spoon!

Slides