Plenary Lecture by Professor Adi Shamir at GYSS 2024

Описание к видео Plenary Lecture by Professor Adi Shamir at GYSS 2024

“Why are adversarial examples so close to the original, and why don’t they resemble the target class?” Professor Adi Shamir (2002 Turing Award recipient) questioned. To address the problem statement, Prof Shamir introduced his intuitive Dimpled Manifold Model during his plenary lecture, which describes 2D decision boundaries in a 3D space. A better understanding of such decision-making processes will shed new light on deep neural networks.

Комментарии

Информация по комментариям в разработке

Похожие видео