Fengshuo Bai
 |
PhD. Student
|
Biography
I am currently a Ph.D. Student at Department of Computer Science and Engineering, Shanghai Jiao Tong University as well as a member of PAIR-Lab, co-advised by Prof. Yaodong Yang and Prof. Ying Wen. I am also selected into Wenjun Wu Honored Ph.D. Class in 2023, advised by Prof. Cewu Lu. My research interest lies in Dexterous Manipulation, Preference-based RL and AI Alignment. If you would like to discuss about potential collaboration or common research interests, please do not hesitate to contact me.
News 🔥
📢 2025.9 Our paper, “ValueDCG,” has been accepted to the NeurIPS 2025 Workshop on Regulatable ML (RegML).
📢 2025.9 Two papers were accepted to NeurIPS’25, with one selected as Spotlight!
📢 2025.8 Our paper has been accepted at Conference on Empirical Methods in Natural Language Processing (EMNLP’25)!
📢 2025.7 We have released our VLA survey titled “A Survey on Vision-Language-Action Models: An Action Tokenization Perspective.”
📢 2025.6 Our paper has been accepted for oral presentation at the Artificial General Intelligence Conference (AGI’25)!
Selected Papers
(* indicates equal contribution)
STAR: Efficient Preference-based Reinforcement Learning via Dual Regularization
Fengshuo Bai, Rui Zhao, Hongming Zhang, Sijia Cui, Shao Zhang, Ying Wen, Yaodong Yang, Bo Xu, Lei Han
Conference on Neural Information Processing Systems (NeurIPS), 2025
A Scalable and Self-improving Data Generation Framework for Dexterous Manipulation Spotlight
Kefei Zhu, Fengshuo Bai , YuanHao Xiang, Yishuai Cai, Xinglin Chen, Ruochong Li, Xingtao Wang, Yaodong Yang, Hao Dong, Xiaopeng Fan, Yuanpei Chen
Conference on Neural Information Processing Systems (NeurIPS), 2025
ValueDCG: Measuring Comprehensive Human Value Understanding Ability of Language Models
Zhaowei Zhang, Fengshuo Bai*, Jun Gao*, Yaodong Yang
Workshop on Neural Information Processing Systems (NeurIPS), Regulatable ML 2025
AdaptFlow: Adaptive Workflow Optimization via Meta-Learning
Runchuan Zhu, Bowen Jiang, Lingrui Mei, Fangkai Yang, Lu Wang, Haoxiang Gao, Fengshuo Bai, Pu Zhao, Qingwei Lin, Saravan Rajmohan, Dongmei Zhang
Conference on Empirical Methods in Natural Language Processing (EMNLP), 2025
A Survey on Vision-Language-Action Models: An Action Tokenization Perspective
Yifan Zhong*, Fengshuo Bai*, Shaofei Cai, Xuchuan Huang, Zhang Chen, Xiaowei Zhang, Yuanfei Wang, Shaoyang Guo, Tianrui Guan, Ka Nam Lui, Zhiquan Qi, Yitao Liang, Yuanpei Chen, Yaodong Yang
Roadmap on Incentive Compatibility for Al Alignmentand Governance in Sociotechnical Systems Oral
Zhaowei Zhang, Fengshuo Bai, Mingzhi Wang, Haoyang Ye, Chengdong Ma, Yaodong Yang
Artificial General Intelligence Conference (AGI), 2025
Retrieval Dexterity: Efficient Object Retrieval in Clutters with Dexterous Hand
Fengshuo Bai, Yu Li, Jie Chu, Tawei Chou, Runchuan Zhu, Ying Wen, Yaodong Yang, Yuanpei Chen
Amulet: ReAlignment During Test Time for Personalized Preference Adaptation of LLMs
Zhaowei Zhang*, Fengshuo Bai*, Qizhi Chen, Chengdong Ma, Mingzhi Wang, Haoran Sun, Zilong Zheng, Yaodong Yang
International Conference on Learning Representations (ICLR), 2025
GRAIT: Gradient-Driven Refusal-Aware Instruction Tuning for Effective Hallucination Mitigation
Runchuan Zhu, Xinke Jiang, Jiang Wu, Zhipeng ma, Jiahe Song, Fengshuo Bai, Dahua Lin, Lijun Wu, Conghui He
Annual Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics (NAACL), 2025
β-DQN: Improving Deep Q-Learning By Evolving the Behavior Oral
Hongming Zhang, Fengshuo Bai, Chenjun Xiao, Chao Gao, Bo Xu, Martin Müller
International Conference on Autonomous Agents and Multiagent Systems (AAMAS), 2025
RAT: Adversarial Attacks on Deep Reinforcement Agents for Targeted Behaviors Oral
Fengshuo Bai, Runze Liu, Yali Du, Ying Wen, Yaodong Yang
Proceedings of the AAAI Conference on Artificial Intelligence (AAAI), 2025
Efficient Model-agnostic Alignment via Bayesian Persuasion
Fengshuo Bai, Mingzhi Wang, Zhaowei Zhang, Boyuan Chen, Yinda Xu, Ying Wen, Yaodong Yang
PEARL: Zero-shot Cross-task Preference Alignment and Robust Reward Learning for Robotic Manipulation
Runze Liu, Yali Du, Fengshuo Bai, Jiafei Lyu, Xiu Li
Conference on International Conference on Machine Learning (ICML), 2024
PiCor: multi-task deep reinforcement learning with policy correction Oral
Fengshuo Bai, Hongming Zhang, Tianyang Tao, Zhiheng Wu, Yanna Wang, Bo Xu
Proceedings of the AAAI Conference on Artificial Intelligence (AAAI), 2023
Meta-reward-net: Implicitly differentiable reward learning for preference-based reinforcement learning
Runze Liu, Fengshuo Bai, Yali Du, Yaodong Yang
Conference on Neural Information Processing Systems (NeurIPS), 2022
Experiences
Institute for AI, Peking University (Jan 2022 - Oct 2022)
Research on algorithms of Reinforcement Learning from the Human Feedback.
Supervisor: Dr. Yali Du, Dr. Yaodong Yang
Honors & Awards
Pacemaker to Merit Student, 2018.09
President Scholarship of Southeast University, 2018.09
China National Scholarship, 2017.09
Competitions
Sliver Prize (25/1178), Kaggle Lux AI Challenge, 2021.12
Third (Intro), 7th (Research) NeurIPS 2021 MineRL Diamond Competition, 2021.10
14th, IJCAI 2020 Mahjong Artificial Intelligence Competition, 2021.01
Services
I was a reviewer / PC member of conferences:
DAI 2023-2024
AAMAS 2024-2026
NeurIPS 2024-2025
ICLR 2025-2026
AAAI 2025-2026
ICML 2025
IJCAI 2025
IROS 2025