Research directions include, but are not limited to:

Foundation Models

• Generalization algorithms based on reinforcement learning

• Physically interpretable model mechanism analysis and architecture design

Multimodal Understanding and Representation

• Image/video understanding and controllable generation

• Language model-augmented proteomics analysis

Human-Machine / Multi-Agent Collaboration

• Robot control and computer operation driven by vision-language models

• Multi-agent collaborative reasoning based on reinforcement learning

AI for Science and Innovation

• AI-driven hypothesis generation and frontier theoretical validation in biology, chemistry, physics, and mathematics