World Modelling in Embodied AI Developing structured world models to enable more robust, adaptive, and data-efficient embodied intelligence, with autonomous driving as a core testbed Knowledge-Driven Agentic System for Science Developing AI agents that leverage knowledge for scientific reasoning and discovery Autonomous Intelligence Modeling Enhancing deep reasoning and systematic generalization in large models through an Autonomous Intelligence Framework that integrates symbolic rules and data. Human–AI Co‑Evolution and Agentic Research Exploring a new research and education paradigm built around human–AI co‑evolution and agentic collaboration
