Bohan Zhou 周伯涵

Embodied Intelligence · Foundation Models

I'm currently a Ph.D. student (2026 - 2030) in the Department of Computer Science and Engineering at The Chinese University of Hong Kong (CUHK), supervised by James Cheng. My current research focuses on embodied intelligence and foundation models, with interests in multimodal learning, reinforcement learning and generalizable agents. Previously, I received my master's degree from Peking University (2023 - 2026), supervised by Zongqing Lu, and my bachelor's degree from Nankai University (2019 - 2023), supervised by Jingtai Liu. Feel free to drop me an email to discuss or share wonderful ideas!

News

Nov 13, 2026

AtmosBench / PhysicalSDC is accepted by KDD 2027, focusing on physics-aligned atmospheric backscatter retrieval from spaceborne lidar! ✨

Oct 14, 2026

GroundingPI reached Hugging Face Daily Papers #1, focusing on grounding for physical intelligence! ✨

Jan 13, 2026

DARA is accepted by WWW 2026, focusing on LLM RL post-training for advertising! ✨

Nov 08, 2025

BiDexHD is accepted by AAAI 2026 (oral), focusing on generalizable bimanual dexterous manipulation! ✨

Sep 19, 2025

MEgoHand is accepted by NeurIPS 2025, focusing on hand-object interaction motion generation! ✨

Jun 16, 2025

NOLO is accepted by IROS 2025 (oral), focusing on offline RGB video navigation! ✨

May 01, 2025

Cradle v2.0 is accepted by ICML 2025 (Cradle v1.0 is accepted by ICLR 2024 Workshop and NeurIPS 2024 Workshop),
focusing on general computer control (LMMagent)! ✨

Jan 22, 2025

CrossDex is accepted by ICLR 2025, focusing on cross-embodiment dexterous grasping! ✨

Jan 20, 2025

AERO is accepted by WWW 2025, focusing on reinforcement learning for blockchain! ✨

Jul 01, 2024

UniCode is accepted by ECCV 2024, focusing on general multimodal representation! ✨

Jul 01, 2024

PVDR is accepted by ECCV 2024, focusing on pre-trained representation for policy learning! ✨

Sep 22, 2023

STG is accepted by NeurIPS 2023, focusing on learning from visual observations! ✨

Mar 04, 2023

GFIE is accepted by CVPR 2023, focusing on human-object interaction (benchmark)! ✨

Publications

Preprints & Under Review

RoboAware: state-locked counterfactual search, execution-aware Q-learning and closed-loop deployment

RoboAware: Learning to Coordinate Embodied Skills from Counterfactual Outcomes

Bohan Zhou, et al. (2026)
FETCH overview: evidence distillation, FETCH-7B and FETCH-Bench

FETCH: Frame-grounded Evidence Topology for Embodied Spatial Planning with the Counterfactual Hypothesis

Bohan Zhou, et al. (2026)
GroundingPI research preview

GroundingPI: A Grounding Foundation Model towards Physical Intelligence with Visual Primitives

Qize Yu, ..., Bohan Zhou, et al. (2026)
Being-0 research preview

Being-0: A Humanoid Robotic Agent with Vision-Language Models and Modular Skills

Haoqi Yuan, Yu Bai, Yuhui Fu, Bohan Zhou, Yicheng Feng, Xinrun Xu, Yi Zhan, Börje F. Karlsson, Zongqing Lu (2025)

Published & Accepted Papers

Atmospheric backscatter retrieval comparison with PhysicalSDC and baseline methods

The Floor Is Not the Limit: Recovering Atmospheric Backscatter from Spaceborne Lidar via Physics-Aligned Encoding

Tianchi Xu, Yuanpeng He, Jintao Xu, Kang He, Linyu Li, Huanyao Zhang, Dongming Jin, Bohan Zhou, Wenpin Jiao, Zhi Jin
KDD 2027
AffordanceVLA overview: progressive data curriculum, model architecture and evaluation

AffordanceVLA: A Vision-Language-Action Model Empowering Action Generation through Affordance-Aware Understanding

Qize Yu, ..., Bohan Zhou, et al.
ECCV 2026 Workshop
DARA: Few-shot Budget Allocation in Online Advertising via In-Context Decision Making with RL-Finetuned LLMs

DARA: Few-shot Budget Allocation in Online Advertising via In-Context Decision Making with RL-Finetuned LLMs

Mingxuan Song, Yusen Huo, Bohan Zhou, Shenglin Yin, Zhen Xiao*, Jieyi Long, Zhilin Zhang, Chuan Yu*
WWW 2026 oral
Cradle: Empowering Foundation Agents towards General Computer Control

Cradle: Empowering Foundation Agents towards General Computer Control

Weihao Tan, Wentao Zhang, Xinrun Xu, Haochong Xia, Ziluo Ding, Boyu Li, Bohan Zhou, Junpeng Yue, Jiechuan Jiang, Yewen Li, Ruyi An, Molei Qin, Chuqiao Zong, Longtao Zheng, YuJie Wu, Xiaoqiang Chai, Yifei Bi, Tianbao Xie, Pengjie Gu, Xiyun Li, Ceyao Zhang, Long Tian, Chaojie Wang, Xinrun Wang, Börje F. Karlsson, Bo An, Shuicheng Yan, Zongqing Lu
ICML 2025
AERO: Enhancing Sharding Blockchain via Deep Reinforcement Learning for Account Migration

AERO: Enhancing Sharding Blockchain via Deep Reinforcement Learning for Account Migration

Mingxuan Song, Pengze Li, Bohan Zhou, Shenglin Yin, Zhen Xiao, Jieyi Long
WWW 2025 oral
UniCode: Learning a Unified Codebook for Multimodal Large Language Models

UniCode: Learning a Unified Codebook for Multimodal Large Language Models

Sipeng Zheng, Bohan Zhou, Yicheng Feng, Zongqing Lu
ECCV 2024
Pre-trained Visual Dynamics Representations for Efficient Policy Learning

Pre-trained Visual Dynamics Representations for Efficient Policy Learning

Hao Luo, Bohan Zhou, Zongqing Lu
ECCV 2024
Towards General Computer Control: A Multi Modal Agent For Red Dead Redemption II As A Case Study

Towards General Computer Control: A Multi Modal Agent For Red Dead Redemption II As A Case Study

Weihao Tan, Ziluo Ding, Wentao Zhang, Boyu Li, Bohan Zhou, Junpeng Yue, Haochong Xia, Jiechuan Jiang, Longtao Zheng, Xinrun Xu, Yifei Bi, Pengjie Gu, Xinrun Wang, Börje F. Karlsson, Bo An, Zongqing Lu
ICLR / NeurIPS 2024 Workshop
GFIE: A Dataset and Baseline for Gaze-Following from 2D to 3D in Indoor Environments

GFIE: A Dataset and Baseline for Gaze-Following from 2D to 3D in Indoor Environments

Zhengxi Hu, Yuxue Yang, Xiaolin Zhai, Dingye Yang, Bohan Zhou, Jingtai Liu
CVPR 2023
Gaze Target Estimation Inspired by Interactive Attention

Gaze Target Estimation Inspired by Interactive Attention

Zhengxi Hu , Kunxu Zhao , Bohan Zhou, Hang Guo , Shichao Wu , Yuxue Yang, Jingtai Liu
TCSVT 2022

Education

Department of Computer Science and Engineering, The Chinese University of Hong Kong
Ph.D. in Computer Science and Engineering (2026.08 - Present).
Research: embodied intelligence and foundation models.
Advisor: Prof. James Cheng.
National Engineering Research Center of Visual Technology, Peking University
Master's degree in Computer Applied Technology (2023.09 - 2026.07).
Research: multimodal learning, reinforcement learning and agents. GPA: 3.88/4.0.
Advisor: Prof. Zongqing Lu.
College of Artificial Intelligence, Nankai University
Bachelor's degree in Intelligent Science and Technology (2019.09 - 2023.07).
GPA: 3.95/4.0; rank: 2/98. Recipient of seven undergraduate scholarships.
Advisor: Prof. Jingtai Liu.

Experience

KnowinAI
Robot Deployment and Embodied Skill Coordination.
Delivered Astra simulation/real-robot evaluations and dual-arm dexterous systems, integrating perception, control and hardware. Led RoboAware, learning state-dependent coordination from same-state counterfactual branches with MCTS and Q-learning while keeping low-level executors frozen.
BeingBeyond
Embodied Foundation Models and Dexterous Systems.
Contributed the human-video data pipeline for Being-H0, combining MEgoHand, TASTE-Rob and EgoDex for robot-policy transfer. Helped develop RealDexUMI, a wearable dexterous data-collection system with isomorphic control and integrated vision-tactile end effectors.
Momenta
M-star Talent Program.
Built a scenario-mining and cleaning loop, raising intersection-action recognition from about 70% to 97% and reducing annotation costs by 51%–62%. Coordinated R6 post-training at the 128-GPU scale and led Flow Matching trajectory generation with FPO safety enhancement, with 30 km of IM LS6 road testing at >10 Hz; the approach entered current products.
JD Explore Academy
TGT Talent Program.
Helped build EgoLive: 1,680 hours, 65K episodes and 346 manipulation tasks in unconstrained real environments, with stereo RGB/IMU, hand reconstruction, 3D reconstruction and semantic annotation. Transferred MEgoHand pretraining to VLA-H2 dexterous policies through object modeling, cross-embodiment retargeting and trajectory optimization.
Beijing Academy of Artificial Intelligence (BAAI)
Research Intern · Generalist and Embodied Agents.
Worked on video-driven policy learning, visual navigation and multimodal dexterous manipulation: STG, NOLO, MEgoHand and BiDexHD. Produced four first-author papers; STG, NOLO and MEgoHand were featured by Prof. Zongqing Lu at the BAAI Conference in three consecutive years, 2023–2025.

Awards & Honors

3rd Prize, Gripper Track, the third X Square–Tsinghua Embodied AI Challenge.
Beijing Outstanding Graduate, Peking University, 2026.
National Scholarship; Outstanding Student Award, Peking University, 2025.
Tianjin Municipal Outstanding Graduate, Nankai University, 2023.
Tianjin Municipal People's Government Scholarship (top 2%), 2021.
Gongneng Scholarship (top 5%), Nankai University, 2020 and 2022.
2nd Prize, National College Student Mathematical Modeling Competition, 2021.
Honorable Mention, American College Mathematical Contest in Modeling, 2021.

Services

Teaching Assistant:
  • Introduction to Computing, 2025 Fall, Prof. Rui Gan, Department of Mathematics, Peking University.
  • Fundamental Computing with C++, 2026 Fall, Dr. Yat Chiu Law, Senior Lecturer, Department of Computer Science and Engineering, The Chinese University of Hong Kong.

Reviewer:
ICLR 2025, ICML 2025, NeurIPS 2025, AAAI 2026, AAAI 2027, ICLR 2027

Co-authors & Buddies