Peter Kairouz

Redirection for Erasing Memory (REM): Towards a universal unlearning method for corrupted data

Differentially Private Conditional Text Generation with RL-Boosted Control

Can Large Language Models Really Recognize Your Name?

Should We Forget About Certified Unlearning? Evaluating the Pitfalls of Noisy Methods

Language Models May Verbatim Complete Text They Were Not Explicitly Trained On

URANIA: Differentially Private Insights into AI Use

iFedDR: Auto-Tuning Local Computation with Inexact Douglas-Rachford Splitting in Federated Learning