AI theory, machine learning, safety and alignment of large models, generative AIs, etc.

Developing learning systems that rapidly acquire knowledge and adapt to new environments

The Convergence and Empowerment of AI and Neuroscience, and Brain-Computer Interface Algorithms

Advancing scientific observation with AI to accelerate scientific discovery

The cross-empowerment between artificial intelligence and software engineering: AI for SE and SE for AI

Egocentric Vision & Multimodal Generative AI

Mainly engaged in research related to multimodal large models and natural language processing.

Interactive Embodied Intelligence powered by World Model, Dexterous Control and Sensing, and Embodied Foundation Model.

Structured Representation Learning for Scientific Intelligence: Bidirectional Synergy between AI and Science

(Multi-)Agent and Their Economics

AI+Science

We develop Physics-informed AI for Science, targeting astronomical imaging, atmospheric turbulence sensing, and computational optics to push observational detection limits beyond classical physics and close the loop from raw photons to astrophysical disco