Sean Geiger
Papers
1
Total Citations
5
H-Index
1
About
Sean Geiger is a researcher advancing the frontiers of machine learning, with a primary focus on imitation learning and reinforcement learning. His key contributions lie in developing sample-efficient algorithms for learning complex behaviors from observation alone, a critical challenge for robotics and autonomous systems. Geiger is best known for his influential work on adversarial imitation learning from observation, where he pioneered methods that enable agents to master tasks by watching state-only demonstrations, bypassing the need for costly action labels. His 2019 paper on "Sample-efficient Adversarial Imitation Learning from Observation" (5 citations) introduced a novel framework that dramatically reduces the data required for imitation, achieving state-of-the-art performance on complex continuous control tasks. This work has been recognized for its potential to make imitation learning practical for real-world applications where expert action data is unavailable. Geiger's research continues to push boundaries in sample-efficient learning, with his methods offering a promising path toward more autonomous and adaptable AI systems that can learn from passive observation.
Research Focus
Key Achievements
Top Papers
- 1Sample-efficient Adversarial Imitation Learning from Observation5 citations · 2019