William Raffe
Papers
1
Total Citations
7
H-Index
1
About
William Raffe is a researcher whose work sits at the intersection of reinforcement learning, games, and human-computer interaction. His most-cited paper, “Learning Options From Demonstrations: A *Pac-Man* Case Study” (2017, 7 citations), tackles a fundamental challenge in RL: the inefficiency of learning from scratch. By introducing a method that learns hierarchical “options” from human demonstrations, Raffe’s work shows how agents can bypass costly trial-and-error phases and adopt more intelligent, structured behaviors—using the classic game of *Pac-Man* as a compelling testbed. This contribution is particularly valuable for applications in games, robotics, and control systems, where rapid, safe learning is critical. Though his citation count is modest, Raffe’s focus on bridging demonstration-based learning with hierarchical RL offers a practical pathway toward more sample-efficient and interpretable AI. His research speaks directly to students and practitioners seeking to build agents that learn faster and more naturally from human guidance, making his work a thoughtful step forward in making reinforcement learning more accessible and effective in real-world scenarios.
Research Focus
Key Achievements
Top Papers
- 1