Ulrich Aïvodji

Biography

Ulrich Aïvodji studies what an automated decision reveals about the people behind its training data, and whether the reasons a system gives for that decision can be trusted. His work on fairwashing shows how an explanation can make a discriminatory model appear fair. He develops ways to audit these claims and test privacy protections in machine learning.

Research topics

• Algorithmic fairness
• Model explanations
• Privacy auditing