This block seminar focuses on physically realizable adversarial examples. We will examine recently published attacks on deep learning algorithms and discuss their impact on real-world systems. We will also look at possible defenses and countermeasures to protect the systems against these attacks. During the course of the semester, you will conduct your own literature review, summarize your findings in a short paper, and present these results in a lightning talk as well as during a poster session. We're mimicking our own little conference!
This course is meant for Master students.
- Trainer/in: Pia Hanfeld
- Trainer/in: Konrad Rieck