Kianté Brantley

Research:

I am an Assistant Professor of Computer Science at Kempner Institute and School of Engineering and Applied Sciences (SEAS), Harvard University. My research focuses on problems at the intersection of machine learning and interactive decision-making, with the goal of improving the decision-making capabilities of foundation models. This involves studying, building, and improving techniques in reinforcement learning, imitation learning, and natural language processing.

Prospective Students:

I am recruiting PhD students this cycle! I am looking for candidates with a strong background in applied reinforcement learning, natural language processing, machine learning, and deep learning. If you are applying to [Harvard], please mention my name in your research statement. For those who have been admitted to the Harvard PhD program and would like to work with me, please send your CV along with a brief overview of your research interests.

Recent Publications:

Regressing the Relative Future: Efficient Policy Optimization for Multi-turn RLHF

Gao, Zhaolin and Zhan, Wenhao and Chang, Jonathan D and Swamy, Gokul and Brantley, Kianté and Lee, Jason D and Sun, Wen

Preprint, 2024

[Arxiv]

REBEL: Reinforcement Learning via Regressing Relative Rewards

Gao, Zhaolin and Chang, Jonathan D and Zhan, Wenhao and Oertell, Owen and Swamy, Gokul and Brantley, Kianté and Joachims, Thorsten and Bagnell, J Andrew and Lee, Jason D and Sun, Wen

Preprint, 2024

[Arxiv]

Dataset Reset Policy Optimization for RLHF

Chang, Jonathan D and Shan, Wenhao and Oertell, Owen and Brantley, Kianté and Misra, Dipendra and Lee, Jason D and Sun, Wen

Preprint, 2024

[Arxiv]

Coactive Learning for Large Language Models using Implicit User Feedback

Tucker, Aaron David and Brantley, Kianté and Cahall, Adam and Joachims, Thorsten

Proceedings of the International Conference on Machine Learning (ICML), 2024

[Arxiv]