Lightweight World Model and Planning for Autonomous Driving
Research on planning-oriented world modeling with multi-camera BEV perception, ConvLSTM temporal modeling, and staged end-to-end training.
Research systems, agents, and applied AI builds.
Research on planning-oriented world modeling with multi-camera BEV perception, ConvLSTM temporal modeling, and staged end-to-end training.
A video-driven symbolic simulator and belief-state planning benchmark built from HD-EPIC egocentric cooking videos.
A legal QA and agent system integrating Graph-RAG, reinforcement learning, harness engineering, and multi-agent collaboration.
An LLM-agent direction for material design, formulation-performance prediction, and scientific workflow support.