Trustworthy Machine Learning

Semester: Summer 2025

Added: Apr 16, 2025

Summer 2025

Instructor: Adam Dziedzic and Franziska Boenisch

Contact: adam.dziedzic@cispa.de, franziska.boenisch@cispa.de

Lectures: Wednesday, Video-based lectures with interactive Q&A.

CMS

Description

The deployment of machine learning applications in real-world systems necessitates methods to ensure their trustworthiness. This course explores the different aspects of trustworthy machine learning, including Privacy, Collaborative Learning, Model Confidentiality, Robustness, Fairness and Bias, Explainability, Security, and Governance.

Assignments

The course entails 4 practical graded assignments to be handled in groups of two:

  • Privacy: Implement a membership inference attack. Due 28.5.
  • Model extraction: Extract model behavior over an API. Due 25.6.
  • Robustness: Train a model robust against adversarial examples. Due 9.7.
  • Fairness: Train a classifier with high demographic parity. Due 23.7.

Schedule

Date Topic Optional Readings
16.04. Overview, Administration & Intro
30.04. Privacy Part I
07.05. Privacy Part II
28.04. Model stealing and defenses (SL and SSL)
04.06. Midterm-Exam (2:00 PM - 2:45 PM, Location: HS I in E 2.5) Q&A on Robustness (not in Midterm)
11.06. Adversarial Machine Learning / Robustness
18.06. Collaborative learning (18.6.a, 18.6.b)
25.06. Fairness and bias
02.07. Explainability
09.07. Security and Governance / Summary & Open Questions
30.07. Final Exam (2:00 PM - 4:00 PM, Location: HS I in E 2.5)

Feedback

You can ask your questions in the respective thread in the Forum on CMS.

Course Staff