Huy Thuc Ha

Papers

Total Citations

H-Index

About

Huy Thuc Ha is a leading researcher in embodied AI and robot learning, whose work bridges the critical gap between 2D vision-language models and real-world 3D robotic manipulation. Ha’s research centers on three key areas: open-world 3D scene understanding, scalable robot skill acquisition, and complex deformable object manipulation. In his highly influential work “Semantic Abstraction” (2022, 21 citations), Ha pioneered a framework that enables robots to reason about unstructured 3D environments using open-set vocabularies, a foundational capability for robots operating beyond controlled lab settings. His follow-up work, “Scaling Up and Distilling Down” (2023, 12 citations), introduced a novel pipeline that leverages large language models to massively scale data generation for robot training, then distills this data into robust, multi-task visuo-motor policies—significantly advancing the practicality of language-conditioned robotics. Ha also tackled one of robotics’ hardest challenges in “Bag All You Need” (2023, 11 citations), developing a generalizable system for heterogeneous bagging that handles complex interactions between rigid and deformable objects under severe occlusions. With a growing citation footprint and work that directly addresses the core bottlenecks of deploying robots in the wild, Ha is shaping the future of generalist robotic systems.