Published multiple papers at the AAAI Conference on Artificial Intelligence, such as 'The Perils of Trial-and-Error Reward Design: Misdesign through Overfitting and Invalid Task Specifications' (2023), 'Learning Optimal Advantage from Preferences and Mistaking it for Reward' (2024), etc.; Developed Bayes-TrEx tool to improve model transparency.
Research Experience
Was a AAAS AI Policy Fellow in the U.S. Senate; conducts research on human-AI/robot interactions, reinforcement learning, and AI policy.
Background
CS Prof at Brown University. Research interests include human-AI/human-robot interaction, reinforcement learning, and AI Policy.
Miscellany
Interested in effective communication of logical statements; open to external applicants for PhD programs or postdoc positions.