Papers
15
Total Citations
395
H-Index
8
About
Jan Hendrik Metzen is a leading researcher at the intersection of robotics and machine learning, with a primary focus on enabling intelligent, autonomous robotic manipulation. His most impactful work, a 2015 survey on surgical robotics (238 citations), critically examines how machine learning can move surgical systems beyond mere enhanced dexterity toward truly intelligent and autonomous actions. Metzen has made foundational contributions to skill learning, particularly through his work on contextual policy search and active learning. His 2014 paper on "Active contextual policy search" (30 citations) addresses the challenge of learning versatile skills by allowing an agent to actively select which tasks to practice, while his "Minimum Regret Search" (2016) introduces a novel Bayesian optimization acquisition function that balances exploration and exploitation. Metzen also developed the BesMan Learning Platform (2018, 16 citations), a stand-alone system for automated robot skill learning adaptable to various platforms. His research extends to human-robot interaction, including unsupervised segmentation of human movement (16 citations) and intuitive robot interfaces. With a career spanning over a decade, Metzen's work bridges theoretical advances in machine learning with practical robotic applications, particularly in surgical and manipulation contexts, establishing him as a key figure in the drive toward more capable and autonomous robotic systems.
Research Focus
Key Achievements
Top Papers
- 1
- 2Active contextual policy search30 citations · 2014
- 3Towards Learning of Generic Skills for Robotic Manipulation21 citations · 2013
- 4Intuitive Interaction with Robots – Technical Approaches and Challenges18 citations · 2015
- 5The BesMan Learning Platform for Automated Robot Skill Learning16 citations · 2018
- 6
- 7Minimizing Calibration Time for Brain Reading13 citations · 2011
- 8Minimum Regret Search for Single- and Multi-Task Optimization9 citations · 2016
- 9
- 10