Yandan Yang

Beijing Academy of Artificial Intelligence

Papers

Total Citations

H-Index

About

Yandan Yang is a rising researcher at the intersection of computer vision, embodied AI, and 3D scene understanding. Her work centers on bridging the gap between virtual simulation and physical reality, with a particular focus on automated 3D scene generation for training intelligent agents. Her most-cited paper, "MetaScenes: Towards Automated Replica Creation for Real-world 3D Scans" (2025), tackles a critical bottleneck in embodied AI research: the scarcity of high-quality, diverse 3D scenes that faithfully replicate real-world object complexity. By proposing a framework for automated replica creation, Yang addresses the fundamental challenge of sim-to-real transfer, enabling more effective skill acquisition and generalization for embodied agents. Her contributions are especially significant given that existing datasets often fail to capture the nuanced diversity of physical environments. Though early in her career, Yang's work has already garnered attention for its practical implications in robotics and virtual training. Her research promises to accelerate progress in embodied AI by providing the scalable, realistic 3D data infrastructure necessary for developing truly generalizable intelligent systems.