Julian Gao
Papers
3
Total Citations
253
H-Index
3
About
Julian Gao is a leading researcher in robot learning, with a focus on enabling machines to acquire complex manipulation skills from limited human demonstrations. His work bridges imitation learning, few-shot learning, and neural program induction to create robots that can generalize across hierarchical tasks. Gao is best known for introducing **Neural Task Programming (NTP)**—a framework that takes a single video demonstration as input and recursively decomposes it into sub-tasks, allowing robots to perform novel variations of a task without retraining. This seminal work, first presented in 2017 and later expanded in 2018, has accumulated over **170 citations** and is widely recognized for advancing task-level generalization in robotics. He also co-developed **RoboTurk**, a crowdsourcing platform that enables scalable collection of high-quality human demonstration data for imitation learning, addressing a critical bottleneck in robot skill acquisition. With over **80 citations**, RoboTurk has become a foundational tool for researchers seeking to build large-scale robotic manipulation datasets. Gao’s contributions are shaping the future of data-efficient, generalizable robot learning—paving the way toward robots that can learn new tasks as intuitively as humans do.
Research Focus
Key Achievements
Top Papers
- 1Neural Task Programming: Learning to Generalize Across Hierarchical Tasks153 citations · 2018
- 2
- 3Neural Task Programming: Learning to Generalize Across Hierarchical Tasks19 citations · 2017