CoRL 2026
ExpertGen: Scalable Sim-to-Real Expert Policy Learning from Imperfect Behavior Priors
Scalable expert policy learning that transfers imperfect behavior priors from simulation to real robots.
Zixing Wang 王自兴
Welcome! I am a final-year Ph.D. candidate in Computer Science at Purdue University, advised by Prof. Ahmed H. Qureshi. Previously, I received my M.S. in Robotics and B.A. in Computer Science from the University of Minnesota, where I worked with Prof. Nikolaos Papanikolopoulos.
My research focuses on robotic manipulation, with an emphasis on world-action models, generative policy steering, and soft-body manipulation.
Zixing is pronounced Zee-Shing.
Updates
Recent research and career news.
“Physics-Conditioned Grasping for Stable Tool Use” was accepted to IROS 2026.
“Efficient Sim-to-Real Transfer of World-Action Models from Synthetic Priors” was accepted to the CVPR 2026 EAI Workshop.
I joined Amazon Fauna Robotics as a summer intern.
“PPGuide: Steering Diffusion Policies with Performance Predictive Guidance” was accepted to ICRA 2026.
I joined the Robotics and AI Institute as a research intern.
I passed my Ph.D. preliminary exam and became a Ph.D. candidate.
Earlier updates
“Dynamic Robot Tool Use with Vision Language Models” was accepted to the RSS 2025 R3 Workshop.
I joined Mitsubishi Electric Research Laboratories as a summer research intern.
“Implicit Physics-aware Policy for Dynamic Manipulation of Rigid Objects via Soft Body Tools” was accepted to ICRA 2025.
“DeRi-IGP” was accepted to IEEE Robotics and Automation Letters.
The short paper “DeRi-Bot” was accepted to the CoRL 2023 Learning for Soft Robots Workshop.
“DeRi-Bot” was accepted to IEEE Robotics and Automation Letters and presented at ICRA 2024.
“Efficient Q-Learning over Visit Frequency Maps” was accepted to IROS 2023.
Work Experience
Industry research in robotics and computer vision.
-
Amazon · Fauna Robotics
Robotics AS II Intern
-
Robotics and AI Institute
Research Intern
-
Mitsubishi Electric Research Laboratories
Research Intern
-
Image Sensing Systems
Computer Vision Research Intern
Publications
12 publications
CVPR 2026 EAI Workshop
Efficient Sim-to-Real Transfer of World-Action Models from Synthetic Priors
The first zero-shot sim-to-real transfer of a world-action model for robotic manipulation, trained with synthetic demonstrations and no real-world demonstrations.
ICRA 2026
PPGuide: Steering Diffusion Policies with Performance Predictive Guidance
Inference-time steering for diffusion policies using learned predictions of task performance.
IROS 2026 · RSS 2025 R3 Workshop
Physics-Conditioned Grasping for Stable Tool Use
Selecting stable grasps by conditioning robot tool use on the physical demands of the task.
ICRA 2025
Implicit Physics-aware Policy for Dynamic Manipulation of Rigid Objects via Soft Body Tools
Learning dynamic manipulation policies that exploit the implicit physics of soft tools.
IEEE RA-L 2025
DeRi-IGP: Manipulating Rigid Objects Using Deformable Objects via Iterative Grasp-Pull
A grasp-pull framework for moving rigid objects indirectly through deformable objects.
IEEE RA-L 2023 · ICRA 2024
DeRi-Bot: Learning to Collaboratively Manipulate Rigid Objects via Deformable Objects
Collaborative manipulation of rigid objects through learned interactions with deformable objects.
arXiv 2023
AnyPose: Anytime 3D Human Pose Forecasting via Neural Ordinary Differential Equations
Continuous-time human pose forecasting with flexible prediction horizons.
arXiv 2018
ECO: Egocentric Cognitive Mapping
Egocentric cognitive mapping for representing and reasoning about an agent’s surroundings.
* Equal contribution · † Equal advising