James Davidson
Google (United States), University of Illinois Urbana-Champaign
Papers
5
Total Citations
274
H-Index
5
About
James Davidson is a leading researcher at the intersection of robotics, reinforcement learning, and autonomous navigation. His work focuses on enabling robots to operate intelligently in complex, long-range environments, with key contributions in hierarchical planning and learning under uncertainty. Davidson’s most influential work, “Cognitive Mapping and Planning for Visual Navigation” (2019), has garnered over 218 citations, introducing a novel approach that combines spatial memory with deep learning to allow agents to navigate unseen spaces efficiently. He also pioneered PRM-RL (2018), a hierarchical method that integrates sampling-based path planning with reinforcement learning, enabling robust long-range navigation by learning short-range, dynamics-aware policies. Earlier, Davidson advanced decision-making under uncertainty with his work on exploiting domain knowledge in POMDP planning (2010), and introduced innovative learning paradigms such as robot adversaries for self-supervised task acquisition (2017) and Modulated Policy Hierarchies (2018) for sparse reward settings. His research consistently bridges theoretical planning algorithms with practical robotic systems, making him a notable figure in modern robotics and AI.
Research Focus
Key Achievements
Top Papers
- 1Cognitive Mapping and Planning for Visual Navigation218 citations · 2019
- 2
- 3Supervision via competition: Robot adversaries for learning tasks16 citations · 2017
- 4
- 5Modulated Policy Hierarchies5 citations · 2018