Can I read Evaluating Backdoor Attacks on DNNs on EtoBox?

Evaluating Backdoor Attacks on DNNs by charlize.liebrand is a document available to read on EtoBox.

What is Evaluating Backdoor Attacks on DNNs about?

The document discusses various types of backdoor attacks on deep neural networks, including visible, invisible, and clean label backdoor attacks. It explains how adversaries can insert triggers into models that cause them to behave maliciously while appearing normal under clean inputs. The presentation highlights the effectiveness of these attacks in real-world scenarios and the challenges in detecting them.

Author

charlize.liebrand

Language

EN