Martijn van Otterlo
Radboud University Nijmegen, Vrije Universiteit Amsterdam, KU Leuven
Papers
9
Total Citations
862
H-Index
8
About
Martijn van Otterlo is a researcher whose work spans reinforcement learning, robot cognition, relational learning, and spatial understanding, making significant contributions at the intersection of machine learning and intelligent systems. His most influential work, a 2012 publication on reinforcement learning, has garnered an impressive 675 citations, establishing him as a prominent voice in the field. Van Otterlo has made particularly notable advances in the study of object affordances in robotics — investigating how robots can learn and leverage action possibilities across multi-object configurations, moving beyond the single-object limitations of earlier approaches. His application of probabilistic and statistical relational learning frameworks, including ProbLog, to robotic manipulation tasks demonstrates a sophisticated integration of structured reasoning with real-world machine learning challenges. Beyond robotics, his work extends to hierarchical image understanding using qualitative spatial relations and relational instance-based learning, as well as spatial role labeling in natural language. He has also contributed to the broader community through workshops bridging perception, action, and communication in interactive systems. Van Otterlo's research consistently seeks to unify cognitive and computational principles, offering valuable tools for students and researchers working on intelligent, adaptive machines.
Research Focus
Key Achievements
Top Papers
- 1Reinforcement Learning675 citations · 2012
- 2
- 3Spatial Role Labeling Annotation Scheme18 citations · 2017
- 4Machine learning for interactive systems and robots16 citations · 2013
- 5
- 6
- 7
- 8
- 9