TA - Fairness, Accountability, Confidentiality and Transparency in AI

Course Description

The objective of this course is understanding the technical aspects of Fairness, Accountability, Confidentiality, Transparency in AI, specifically getting to know current challenges in the area and state-of-the-art algorithms. Students will discuss influential articles in the field and work on reproducing recent approaches, thereby also gaining hands-on experience on the process of scientific research in AI and contribute to good reproducibility practices in the field.

In this course students implement an existing FACT paper in groups of 4. They follow the setup from the Machine Learning Reproducibility Challenge (MLRC), and are encouraged to participate in the challenge by submitting their reproduction code and report.

Responsibilities

• Match MLRC paper to students according to their preferences using a minimal cost assignment algorithm
• Supervise 3-4 groups on their reproduction tasks
• Help students submit to the MLRC challenge
• Grade students on their reproduction effort