Text-driven Human Motion Generation
Generating plausible, controllable human motions from language instructions and high-level intentions for embodied agents and creative applications.
Human-centered AI · VR/AR · Human Motion
Human-Centered Artificial Intelligence Lab (HAI Lab)
We build intelligent systems that understand how people see, move, and interact. Our research connects human motion generation, immersive computing, behavior modeling, and human-centered AI.
Research directions
01
02
03
04
05
06
What we work on
HAI Lab studies computational models and interactive systems that make AI more aware of human bodies, intentions, and contexts. Below highlights our major research directions.
Virtual and Augmented Reality
Building immersive systems that use gaze, pose, and interaction signals to understand users in extended reality environments.
Human Motion Prediction
Forecasting future body movements during daily activities, hand-object interaction, and collaborative tasks.
Gaze and Visual Attention
Modeling where people look, why they look there, and how gaze coordinates with hands, head, body, and surrounding context.
Embodied Human-AI Interaction
Designing AI systems that reason about human bodies and interaction intent across physical, virtual, and mixed environments.
Behavior Modeling for Intelligent Systems
Learning reusable representations from interaction data such as mouse behavior, gaze traces, poses, and multimodal user actions.