Jiayi Liu (刘嘉怡)
Ph.D. Student @ HKUST(GZ)
I am currently a second-year Ph.D. student in Artificial Intelligence at the Hong Kong University of Science and Technology (Guangzhou), supervised by Prof. Junwei Liang. My research interests include Artificial Intelligence and Robotics, specifically focusing on robotic manipulation and trajectory prediction.
Previously, I earned the B.Sc. degree in Applied Statistics from Central University of Finance and Economics in 2020, and the M.Sc. in Computational Statistics and Machine Learning from University College London in 2021, supervised by Prof. Ingemar J. Cox. Subsequently, I worked as a researcher at Tencent, developing deep learning solutions for virtual human.
- [2026.02] [New!] One paper (EgoTraj-Bench) accepted by ICRA 2026! See you in Vienna!
- [2025.10] We announce Super-Mimic, a hierarchical framework for zero-shot robotic imitation from unscripted human videos through procedural intent inference, is on Arxiv!
- [2025.10] We announce EgoTraj-Bench, the first real-world benchmark for robust trajectory prediction under noisy ego-centric observations, is on Arxiv!
- [2025.09] One paper (AGNOSTOS) accepted by NeurIPS 2025!
- [2025.05] We announce AGNOSTOS, a novel simulation benchmark designed to assess zero-shot generalization alongside a new VLA method, further supporting advancements in general-purpose robotic manipulation.
The * denotes equal contribution.
Trajectory Prediction
EgoTraj-Bench: Towards Robust Trajectory Prediction Under Ego-view Noisy Observations
International Conference on Robotics and Automation(ICRA), 2026.
Robotic Manipulation
Exploring the Limits of Vision-Language-Action Manipulations in Cross-task Generalization
Conference on Neural Information Processing Systems(NeurIPS), 2025.
Best Novice of Tencent
Dean’s List with Distinction of UCL
A dedicated hiking lover. A so-so baker. A loyal servant of a blue yummy cat.
