Papers
8
Total Citations
248
H-Index
7
About
Jackie Kay is a leading researcher at the intersection of reinforcement learning and robotics, best known for pioneering work in generalist agents and sim-to-real transfer. Her most celebrated contribution is Gato, a multi-modal, multi-task, multi-embodiment generalist policy that applies large-scale language modeling principles to build a single agent capable of performing hundreds of tasks across diverse environments—a landmark paper with 66 citations. Kay has also made significant advances in self-supervised sim-to-real adaptation for visual robotic manipulation (56 citations), enabling robots to learn from unlabeled real-world data without costly reward engineering. Her research on gentle object manipulation using curiosity-driven deep reinforcement learning (45 citations) addresses the critical challenge of robots interacting safely with fragile objects. Additionally, she has developed robust reinforcement learning frameworks for continuous control under model misspecification (38 citations), enhancing the reliability of learned policies in unpredictable real-world conditions. With over 200 total citations across her highly cited works, Kay’s contributions are shaping the future of generalist robotics and safe, data-efficient robot learning.
Research Focus
Key Achievements
Top Papers
- 1A Generalist Agent66 citations · 2022
- 2Self-Supervised Sim-to-Real Adaptation for Visual Robotic Manipulation56 citations · 2020
- 3
- 4
- 5
- 6Learning Dexterous Manipulation from Suboptimal Experts12 citations · 2020
- 7Self-Supervised Sim-to-Real Adaptation for Visual Robotic Manipulation9 citations · 2019
- 8