Papers
7
Total Citations
192
H-Index
4
About
Eric Eaton is a leading researcher in artificial intelligence, specializing in lifelong learning, multi-task learning, and robotics. His work focuses on creating agents that can continuously adapt and improve over time, moving beyond static, single-task models. A key contribution is his pioneering work on *Online Multi-Task Learning for Policy Gradient Methods* (149 citations), which dramatically improves sample efficiency in high-dimensional robotic control by enabling agents to share knowledge across related tasks. Eaton also advances practical, real-world AI through *Lifelong Learning for Disturbance Rejection on Mobile Robots*, demonstrating how robots can adapt to physical changes and degradation. Beyond algorithms, he is deeply committed to AI education, developing interdisciplinary project-driven courses and the widely-used *Gridworld Search and Rescue* simulator to teach integrated AI concepts. His recent work on *Sparse PointPillars* tackles the critical challenge of deploying efficient 3D perception on resource-constrained embedded systems for mobile robots. Through his research, Eaton is shaping a future where intelligent systems are not only more capable, but also more adaptable and enduring.
Research Focus
Key Achievements
Top Papers
- 1Online Multi-Task Learning for Policy Gradient Methods149 citations · 2014
- 2Teaching Integrated AI Through Interdisciplinary Project‐Driven Courses21 citations · 2017
- 3Lifelong learning for disturbance rejection on mobile robots9 citations · 2016
- 4
- 5
- 6
- 7Reports on the 2015 AAAI Workshop Series2 citations · 2015