Selected Topics of Information Security (Security and Privacy of Distributed Learning) (SS 2027)

Content

Distributed learning approaches, such as Federated Learning (FL), enable multiple clients to collaboratively train machine learning models without sharing their private data. With the increasing deployment of distributed learning systems in real-world applications, such as IoT malware detection, mobile risk management, and privacy-sensitive AI applications, ensuring their security, privacy, and reliability becomes increasingly important.

This interactive lecture focuses on security and privacy challenges in distributed learning systems. The course introduces the components of distributed learning systems on both the client and server side, including training procedures and aggregation algorithms. Furthermore, the lecture covers security and privacy attacks and defenses in distributed learning systems, including model poisoning, data poisoning, and backdoor attacks that allow adversaries to manipulate or control the outcome of the learning process. Further topics that will be discussed in the course include robustness mechanisms, privacy-preserving approaches, and security aspects of practical distributed learning deployments.

The course combines theoretical concepts with hands-on exercises of different difficulty levels. Students will gain practical experience by implementing and analyzing selected attacks and defense mechanisms, including state-of-the-art backdoor attacks and mitigation approaches.

Administrative Information

The course will consist of several practical exercises in which students will implement and evaluate different distributed learning concepts, attacks, and defenses. Course-related communication will take place via Discord and TeachCenter. Detailed information on the course organization, exercises, and grading will be provided before the course is offered in the summer term.

Lecture Dates

Date	Begin	End	Location	Event	Type	Comment
2027/03/01	08:15	09:00	CCGEG002	CLASS	VU	CLASS/tbc - Platzhalter-Termin

Lecturers