师资队伍

前沿计算研究中心

王鹤

职称:助理教授

研究所:前沿计算研究中心

研究领域:具身智能,三维视觉,机器人学

办公电话:86-10-6276 1084

电子邮件:hewangpku.edu.cn

个人主页:https://hughw19.github.io/


简介

王鹤博士是北京大学前沿计算研究中心(CFCS)的助理教授和博士生导师。他创立并领导了北大具身感知与交互实验室(EPIC Lab,主页:https://hughw19.github.io),研究目标是通过发展具身技能及具身多模态大模型推进通用具身智能。他同时是北大-银河通用具身智能联合实验室主任,北京智源人工智能研究院具身智能研究中心主任。他已在计算机视觉、机器人学和人工智能的顶级会议和期刊(CVPR/ICCV/ECCV/TRO/RAL/ICRA/NeurIPS/ICLR/AAAI等)上发表五十余篇工作,其论文获得ICCV2023最佳论文候选,ICRA2023最佳操纵论文候选,2022年世界人工智能大会青年优秀论文(WAICYOP)奖,Eurographics 2019最佳论文提名奖。他担任了CVPR2022WACV2022的领域主席,Image and Vision Computing的副主编和诸多顶会的审稿人、程序委员。在加入北京大学之前,他于2021年从斯坦福大学获得博士学位,师从美国三院院士Leonidas. J Guibas教授,于2014年从清华大学获得学士学位。


发表论著

Selected Publications

·        *: equivalent contribution, †: corresponding author

Published Works

https://hughw19.github.io/He_Wang_files/unidexgrasp++.jpg

UniDexGrasp++: Improving Dexterous Grasping   Policy Learning via Geometry-aware Curriculum and Iterative   Generalist-Specialist Learning
 
  Weikang Wan*, Haoran Geng*, Yun Liu, Zikang Shan, Yaodong Yang, Li Yi, He   Wang

 
  ICCV 2023 (
Oral & Best Paper Finalist, final reviews of all   strong accepts)
 
 
arXiv/Project/bibtex

 

 

 

 

 


https://hughw19.github.io/He_Wang_files/gapartnet.png

GAPartNet:   Cross-Category Domain-Generalizable Object Perception and Manipulation via   Generalizable and Actionable Parts
 
  Haoran Geng*, Helin Xu*, Chengyang Zhao*, Chao Xu, Li Yi, Siyuan Huang, He   Wang

 
  CVPR 2023 (
Highlight, final reviews of all accepts)
 
 
arXiv/Project/bibtex

 

 

 

 

 


https://hughw19.github.io/He_Wang_files/3DNav.png

3D-Aware Object   Goal Navigation via Simultaneous Exploration and Identification
 
  Jiazhao Zhang*, Liu Dai*, Fanpeng Meng, Qingnan Fan, Xuelin Chen, Kai   Xu, He Wang

 
  CVPR 2023
 
 
arXiv/Project/bibtex

 

 

 

 

 


https://hughw19.github.io/He_Wang_files/unidexgrasp.png

UniDexGrasp:   Universal Robotic Dexterous Grasping via Learning Diverse Proposal Generation   and Goal-Conditioned Policy
 
  Yinzhen Xu*, Weikang Wan*, Jialiang Zhang*, Haoran Liu*, Zikang Shan, Hao   Shen, Ruicheng Wang, Haoran Geng, Yijia Weng, Jiayi Chen, Tengyu Liu, Li   Yi, He Wang

 
  CVPR 2023
 
 
arXiv/Project/bibtex

 

 

 

 

 


https://hughw19.github.io/He_Wang_files/gapartmanip.png

PartManip:   Learning Cross-Category Generalizable Part Manipulation Policy from Point   Cloud Observations
 
  Haoran Geng*, Ziming Li*, Yiran Geng, Jiayi Chen, Hao Dong, He Wang

 
  CVPR 2023
 
 
arXiv/Project/bibtex

 

 

 

 

 


https://hughw19.github.io/He_Wang_files/so3nf.png

Delving into   Discrete Normalizing Flows on SO(3) Manifold for Probabilistic Rotation   Modeling
 
  Yulin Liu*, Haoran Liu*, Yingda Yin*, Yang Wang, Baoquan Chen
He   Wang
 
  CVPR 2023
 
 
arXiv/SO3NF

 

 

 

 

 


https://hughw19.github.io/He_Wang_files/diga.png

DiGA: Distil to   Generalize and then Adapt for Domain Adaptive Semantic Segmentation
 
  Fengyi Shen, Akhil Gurram, Ziyuan Liu, He Wang
, Alois   Knoll
 
  CVPR 2023
 
 
arXiv/bibtex

 

 

 

 

 


https://hughw19.github.io/He_Wang_files/adazone.png

Adaptive   Zone-aware Hierarchical Planner for Vision-Language Navigation
 
  Chen Gao, Xingyu Peng, Mi Yan, He Wang, Lirong Yang, Haibing Ren,   Hongsheng Li, Si Liu
 
  CVPR 2023
 
  arXiv (soon)

 

 

 


https://hughw19.github.io/He_Wang_files/dexgraspnet.png

DexGraspNet: A   Large-Scale Robotic Dexterous Grasp Dataset for General Objects Based on   Simulation
 
  Ruicheng Wang*, Jialiang Zhang*, Jiayi Chen, Yinzhen Xu, Puhao Li, Tengyu   Liu, He Wang

 
  ICRA 2023 (
Outstanding Manipulation Paper Award Finalist)
 
 
arXiv/Project/bibtex

 

 

 

 

 


https://hughw19.github.io/He_Wang_files/graspnerf.png

GraspNeRF:   Multiview-based 6-DoF Grasp Detection for Transparent and Specular Objects   Using Generalizable NeRF
 
  Qiyu Dai*, Yan Zhu*, Yiran Geng, Ciyu Ruan, Jiazhao Zhang, He Wang

 
  ICRA 2023
 
 
arXiv/Project/bibtex

 

 

 

 

 


https://hughw19.github.io/He_Wang_files/rotlaplace.png

A Laplace-inspired   Distribution on SO(3) for Probabilistic Rotation Estimation
 
  Yingda Yin, Yang Wang, He Wang
, Baoquan Chen
 
  ICLR 2023 (
notable top 25%)
 
 
paper/bibtex

 

 

 

 

 


https://hughw19.github.io/He_Wang_files/se3art.png

Self-Supervised   Category-Level Articulated Object Pose Estimation with Part-Level SE(3)   Equivariance
 
  Xueyi Liu, Ji Zhang, Ruizhen Hu, Haibin Huang, He Wang, Li Yi
 
  ICLR 2023
 
 
paper/bibtex

 

 

 

 

 


https://hughw19.github.io/He_Wang_files/trhoi.png

Tracking and   Reconstructing Hand Object Interactions from Point Cloud Sequences in the   Wild
 
  Jiayi Chen*, Mi Yan*, Jiazhao Zhang, Yenzhen Xu, Xiaolong Li, Yijia Weng, Li   Yi, Shuran Song, He Wang

 
  AAAI 2023 (
Oral Presentation)
 
 
arXiv/Project/bibtex

 

https://hughw19.github.io/He_Wang_files/tro.jpeg

ASRO-DIO: Active   Subspace Random Optimization Based Depth Inertial Odometry
 
  Jiazhao Zhang, Yijie Tang, He Wang, Kai Xu
 
  IEEE Transactions on Robotics (T-RO)
 
 
Paper

 

 


https://hughw19.github.io/He_Wang_files/dreds.png

Domain   Randomization-Enhanced Depth Simulation and Restoration for Perceiving and   Grasping Specular and Transparent Objects
 
  Qiyu Dai*, Jiyao Zhang*, Qiwei Li, Tianhao Wu, Hao Dong, Ziyuan Liu, Ping   Tan, He Wang

 
  ECCV 2022
 
 
arXiv/Project/Code/bibtex

 

 

 

 

 


https://hughw19.github.io/He_Wang_files/maniskill.png

Learning   Category-Level Generalizable Object Manipulation Policy via Generative Adversarial   Self-Imitation Learning from Demonstrations
 
  Hao Shen*, Weikang Wan*, He Wang

 
  Robotics and Automation Letters (RA-L) and IROS 2022
 
 
1st   prize winner of SAPIEN   ManiSkill Challenge 2021 (no external annotation track)
 
 
arXiv/Project/Code/bibtex

 

 

 

 

 


https://hughw19.github.io/He_Wang_files/fishermatch.png

FisherMatch:   Semi-Supervised Rotation Regression via Entropy-based Filtering
 
  Yingda Yin, Yingcheng Cai, He Wang
, Baoquan   Chen
 
  CVPR 2022 (
Oral Presentation)
 
 
arXiv/Project/Code/bibtex

 

 

 

 

 


https://hughw19.github.io/He_Wang_files/rpmg.png

Projective Manifold   Gradient Layer for Deep Rotation Regression
 
  Jiayi Chen, Yingda Yin, Tolga Birdal, Baoquan Chen, Leonidas Guibas, He   Wang

 
  CVPR 2022
 
 
arXiv/Project/Code/bibtex

 

 

 

 

 


https://hughw19.github.io/He_Wang_files/adela.png

ADeLA: Automatic   Dense Labeling with Attention for Viewpoint Adaptation in Semantic   Segmentation
 
  Yanchao Yang*, Hanxiang Ren*, He Wang, Bokui Shen, Qingnan Fan,   Youyi Zheng, C. Karen Liu, Leonidas J. Guibas
 
  CVPR 2022 (
Oral Presentation)
 
 
arXiv/bibtex

 

 

 

 

 


https://hughw19.github.io/He_Wang_files/hoi4d.png

HOI4D: A 4D   Egocentric Dataset for Category-Level Human-Object Interaction
 
  Yunze Liu, Yun Liu, Che Jiang, Kangbo Lyu, Weikang Wan, Hao Shen, Boqiang   Liang, Zhoujie Fu, He Wang, Li Yi
 
  CVPR 2022
 
 
arXiv/bibtex

 

 

 

 

 


https://hughw19.github.io/He_Wang_files/multirobotmap.png

Multi-Robot Active   Mapping via Neural Bipartite Graph Matching
 
  Kai Ye*, Siyan Dong*, Qingnan Fan, He Wang, Li Yi, Fei Xia, Jue   Wang, Baoquan Chen
 
  CVPR 2022
 
 
arXiv/bibtex

 

 

 

 

 


https://hughw19.github.io/He_Wang_files/codedvtr.png

CodedVTR:   Codebook-based Sparse Voxel Transformer with Geometric Guidance
 
  Tianchen Zhao, Niansong Zhang, Xuefei Ning, He Wang, Li Yi, Yu   Wang
 
  CVPR 2022
 
 
arXiv/bibtex

 

 

 

 

 


https://hughw19.github.io/He_Wang_files/dagai.png

Domain Adaptation   on Point Clouds via Geometry-Aware Implicits
 
  Yuefan Shen*, Yanchao Yang*, Mi Yan, He Wang, Youyi Zheng,   Leonidas J. Guibas
 
  CVPR 2022
 
 
arXiv/bibtex

 

 

 

 

 


https://hughw19.github.io/He_Wang_files/esscop.png

Leveraging   SE(3) Equivariance for Self-supervised Category-Level Object Pose Estimation   from Point Clouds
 
  Xiaolong Li, Yijia Weng, Li Yi, Leonidas J. Guibas, A. Lynn Abbott, Shuran   Song, He Wang

 
  NeurIPS 21
 
 
Paper/arXiv/Project/Code/bibtex

 

 

 

 

 


https://hughw19.github.io/He_Wang_files/captra.png

CAPTRA:   CAtegory-level Pose Tracking for Rigid and Articulated Objects from Point   Clouds
 
  Yijia Weng*, He Wang*
, Qiang Zhou, Yuzhe Qin, Yueqi Duan, Qingnan Fan, Baoquan   Chen, Hao Su, Leonidas J. Guibas
 
  ICCV 2021 (
Oral Presentation)
 
 
arXiv/Project/Code/Video/bibtex

 

 

 

 

 


https://hughw19.github.io/He_Wang_files/retrieval.png

Single Image 3D   Shape Retrieval via Cross-Modal Instance and Category Contrastive Learning
 
  Mingxian Lin, Jie Yang, He Wang, Yu-Kun Lai, Rongfei Jia,   Binqiang Zhao, Lin Gao
 
  ICCV 2021
 
 
Paper/Supp./bibtex

 

 

 

 

 


https://hughw19.github.io/He_Wang_files/3dioumatch.png

3DIoUMatch:   Leveraging IoU Prediction for Semi-Supervised 3D Object Detection
 
  He Wang*, Yezhen Cong*, Or Litany, Yue Gao and Leonidas J. Guibas
 
  CVPR 2021
 
 
Paper/Project/Code/Video/bibtex

 

 

 

 

 


https://hughw19.github.io/He_Wang_files/multibodysync.png

MultiBodySync:   Multi-Body Segmentation and Motion Estimation via 3D Scan Synchronization
 
  Jiahui Huang, He Wang, Tolga Birdal, Minkyuk Sung, Federica   Arrigoni, Shi-Min Hu, and Leonidas J. Guibas
 
  CVPR 2021 (
Oral Presentation)
 
 
Paper/Code/Video/bibtex

 

 

 

 

https://hughw19.github.io/He_Wang_files/neuralrouting.png

Robust Neural   Routing Through Space Partitions for Camera Relocalization in Dynamic Indoor   Environments
 
  Siyan Dong*, Qingnan Fan*, He Wang, Ji Shi, Li Yi, Thomas   Funkhouser, Baoquan Chen, Leonidas J. Guibas
 
  CVPR 2021 (
Oral Presentation)
 
 
Paper/Code/bibtex

 

 

 

 

 


https://hughw19.github.io/He_Wang_files/rethink.png

Rethinking   Sampling in 3D Point Cloud Generative Adversarial Networks
 
  He Wang*, Zetian Jiang*, Li Yi, Kaichun Mo, Hao Su, Leonidas J. Guibas
 
  CVPR 2021 Workshop on Learning to Generate 3D Shapes and Scenes
 
 
Paper/Poster/Video/bibtex

 

 

 

 

 


https://hughw19.github.io/He_Wang_files/pt2pc.png

PT2PC: Learning to   Generate 3D Point Cloud Shapes from Part Tree Conditions
 
  Kaichun Mo, He Wang, Li Yi, Xinchen Yan and Leonidas J. Guibas
 
  ECCV 2020
 
 
Paper/Project/Code&Data/bibtex

 

 

 

 

 


https://hughw19.github.io/He_Wang_files/csdf.png

Curriculum DeepSDF
 
  Yueqi Duan*, Haidong Zhu*, He Wang, Li Yi, Ram Nevatia, Leonidas   J. Guibas
 
  ECCV 2020
 
 
Paper/Code/bibtex

 

 

 

 

 


https://hughw19.github.io/He_Wang_files/ancsh.png

Category-level   Articulated Object Pose Estimation
 
  He Wang*, Xiaolong Li*, Li Yi, Leonidas Guibas, A. Lynn Abbott, Shuran   Song
 
  CVPR 2020 (
Oral Presentation)
 
 
Paper/Project/Code&Data/bibtex

 

 

 

 

 


https://hughw19.github.io/He_Wang_files/sapien.png

SAPIEN: A   SimulAted Part-based Interactive ENvironment
 
  Fanbo Xiang, Yuzhe Qin, Kaichun Mo, Yikuan Xia, Hao Zhu, Fangchen Liu,   Minghua Liu, Hanxiao Jiang, Yifu Yuan,
  He Wang, Li Yi, Angel X.Chang, Leonidas J. Guibas and Hao Su
 
  CVPR 2020 (
Oral Presentation)
 
 
Paper/Project/Code&Data/Demo/bibtex

 

 

 

 

 


https://hughw19.github.io/He_Wang_files/posercnn.png

Normalized Object Coordinate   Space for Category-Level 6D Object Pose and Size Estimation
 
  He Wang, Srinath Sridhar, Jingwei Huang, Julien Valentin, Shuran Song,   Leonidas J. Guibas
 
  CVPR 2019 (
Oral Presentation)
 
 
Paper/Project/Code&Data/bibtex

 

 

 

 

 


https://hughw19.github.io/He_Wang_files/gspn.png

GSPN: Generative   Shape Proposal Network for 3D Instance Segmentation in Point Cloud
 
  Li Yi, Wang Zhao, He Wang, Minhyuk Sung, Leonidas Guibas
 
  CVPR 2019
 
 
Paper/Code/bibtex

 

 

 

 

 


https://hughw19.github.io/He_Wang_files/interaction.png

Learning a   Generative Model for Multi-Step Human-Object Interactions from Videos
 
  He Wang*, Soeren Pirk*, Ersin Yumer, Vladimir Kim, Ozan Sener, Srinath   Sridhar, Leonidas J. Guibas
 
  Eurographics 2019 (
Best Paper Honorable Mention)
 
 
Project/Paper/Code/Blog/Video