Aadirupa Saha
@aadirupa.bsky.social
Assistant Professor, UIC.
Research Interests: ML - RL, Online Learning, RLHF, AI Alignment Optimization, Privacy.
Please also connect @AadirupaSaha on X (Twitter)
Website: https://aadirupa.github.io/
Research Interests: ML - RL, Online Learning, RLHF, AI Alignment Optimization, Privacy.
Please also connect @AadirupaSaha on X (Twitter)
Website: https://aadirupa.github.io/
Summary: Agents may game a recommender system with feature manipulation. How do you design systems that resist strategic behavior? Our work bridges online learning & mechanism design, introducing OptGTM to ensure truthful reporting in contextual bandits while minimizing regret.
December 8, 2024 at 10:25 PM
Summary: Agents may game a recommender system with feature manipulation. How do you design systems that resist strategic behavior? Our work bridges online learning & mechanism design, introducing OptGTM to ensure truthful reporting in contextual bandits while minimizing regret.