Bernie Hao Zhu

I am a Ph.D. student in the Paul G. Allen School of Computer Science & Engineering, University of Washington, advised by Prof. Siddhartha Srinivasa. I am affiliated with UW Robotics and Personal Robotics Lab.

I was a researcher at Carnegie Mellon University, advised by Prof. Christopher Atkeson and Prof. Katerina Fragkiadaki. I graduated with High Honors from Zhejiang University, where I earned my bachelor's degree in Computer Science and minored in Electronic Engineering.

My work is TRASH Teaching Robotic Agents Skills of Humans.

Robotics, Computer Vision, Machine Learning
Bimanual robotic manipulation
Contact-rich and dynamics-aware robot learning
Human-robot interaction and alignment

News

[Mar 13, 2024] ADA won the Best Demo Award at HRI 2024. [Nov 03, 2023] I won the UW CSE bakery competition! A photo of me with the giant trophy. [Aug 30, 2023] One paper accepted at CoRL 2023. [Jul 10, 2023] I attended RSS 2023 for our workshop paper. [Jul 12, 2022] My F-1 visa was approved in Consulado General de Estados Unidos en Hermosillo! [Jun 22, 2022] I ended my life as a refugee and returned to the United States. [Feb 13, 2022] I moved to Casablanca, Morocco. This is the first stop of my worldwide wandering. [Feb 23, 2020] One paper accepted at CVPR 2020. [Oct 30, 2019] I attended CoRL 2019 to give a presentation. Huge thanks to organizers for providing wheelchair for me! [Sep 07, 2019] One paper accepted at CoRL 2019. [Jun 16, 2019] I attended CVPR 2019. [Mar 02, 2019] One paper accepted at CVPR 2019.

Selected Publications

Towards General Single-Utensil Food Acquisition with Human-Informed Actions
Ethan K. Gordon*, Amal Nanavati*, Ramya Challa, Bernie Hao Zhu, Taylor A. Kessler Faulkner, Siddhartha S. Srinivasa
CoRL 2023
[Paper][Project][Code]

We discusses the challenges of learning food acquisition policies for robots and proposes a methodology to identify a small subset of actions that can acquire various food items. By mapping human trajectories and clustering, a set of 11 actions is constructed. This set has been shown to have an ≥80% success rate in acquiring different foods, and online learning can be used to determine optimal actions for new food items during a single meal.

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, Hao Su
CVPR 2020, Oral Presentation
[Paper] [Project] [Video]

We build SAPIEN simulator: an interaction-rich and physics-realistic simulation environment integrating PhysX engine and ROS control interface. I lead the team of building SAPIEN dataset: more than 2K 3D articulated models with 14K movable parts. The dataset is richly annotated with kinematic part motions and dynamic interactive attributes to support robot interaction.

S4G: Amodal Single-view Single-Shot SE(3) Grasp Detection in Cluttered Scenes
Yuzhe Qin*, Rui Chen*, Hao Zhu, Meng Song, Jing Xu, Hao Su (*Equal contribution)
CoRL 2019, Spotlight Presentation
[Paper] [Project] [Video] [Presentation]

We studied the problem of 6-DoF grasping by a parallel gripper in a cluttered scene captured using a commodity depth sensor from a single view point. Our learning based approach trained in a synthetic scene can work well in real-world scenarios, with improved speed and success rate compared with state-of-the-arts.

CrowdPose: Efficient Crowded Scenes Pose Estimation and A New Benchmark
Jiefeng Li, Can Wang, Hao Zhu, Yihuan Mao, Hao-Shu Fang, Cewu Lu
CVPR 2019, Oral Presentation
[Paper] [Code] [Dataset] [Media]

We collect CrowdPose, a new dataset of crowded human poses. We propose a joint-candidate single person pose estimation (SPPE) and a global maximum joints association algorithm to tacklethe problem of pose estimation in a crowd. Our method surpasses the state-of-the-art method by 5.2 mAP on CrowdPose and replacing certain steps in the state-of-the-art method with our module would bring 0.8 mAP improvement on MSCOCO.

Miscellaneous

Trivia

My last name Zhu(朱) is pronounced "jew", not "zoo". The signature above includes both my Chinese and English last names. I'm known for my eidetic memory, which consistently gives me an edge in games and tests that require memory :-D I completely lost my hearing in my early 20s. Fortunately, I regained some of it after learning to read lips. Speaking louder to me would be appreciated.

Hobbies

Cooking I cook in a western style, but with Chinese flavors. I'll open my own restaurant one day.

Games I am good at strategy games, board games, poker games, etc.

Music I enjoy classical music and electronic music, especially trance and techno. My faves: I play the piano and suona.