Cheng-Chun Hsu

Papers

Total Citations

H-Index

About

Cheng-Chun Hsu is a robotics researcher focused on bridging the gap between perception and manipulation in everyday environments. His work centers on two key areas: interactive scene understanding and object-centric imitation learning. In his highly regarded 2023 paper “Ditto in the House,” Hsu pioneered methods for robots to build articulation models of indoor scenes through purposeful interaction—enabling machines to understand how doors, drawers, and cabinets move in the real world. This foundational work has already garnered 19 citations and directly supports robot navigation and manipulation in unstructured human spaces. More recently, Hsu introduced SPOT (SE(3) Pose Trajectory Diffusion), an object-centric imitation learning framework that decouples embodiment actions from sensory inputs by representing tasks through SE(3) object pose trajectories relative to targets. This 2025 innovation, with 4 early citations, promises to make robotic learning more generalizable across different hardware platforms. Hsu’s contributions are particularly notable for their practical orientation—his research doesn’t just advance theory but builds systems that can operate in real homes, making him a rising figure in embodied AI and robotic manipulation.