David McAllester
AT&T (United States), Toyota Technological Institute at Chicago
Papers
2
Total Citations
61
H-Index
2
About
David McAllester is a leading figure in artificial intelligence, with foundational contributions spanning reinforcement learning, robotics, and computer vision. His work on the CMUnited-99 robotic soccer system, detailed in his highly cited paper on action selection (38 citations), produced the 1999 RoboCup simulator league champion—a team so dominant that, despite being publicly available for a year, it still placed 4th in the 2000 competition. This research pioneered multi-agent coordination and hierarchical decision-making under real-time constraints. In computer vision, McAllester advanced autonomous navigation with his stereovision-based road boundary detection method (23 citations), which robustly handles challenging environments by combining homography estimation with 3D scene understanding. His broader impact includes seminal theoretical work on PAC-Bayesian learning bounds and structured prediction, which have shaped modern machine learning. With over 10,000 total citations, McAllester’s research continues to influence autonomous systems, from self-driving cars to multi-robot teams, making him a pivotal thinker in bridging theory and practical AI deployment.
Research Focus
Key Achievements
Top Papers
- 1An architecture for action selection in robotic soccer38 citations · 2001
- 2