Jonas Karlsson
Papers
1
Total Citations
97
H-Index
1
About
Jonas Karlsson is a pioneer in the field of artificial intelligence and robotics, best known for his foundational work on hierarchical reinforcement learning and autonomous behavior generation. His seminal 1993 paper, "Learning Multiple Goal Behavior via Task Decomposition and Dynamic Policy Merging," with 97 citations, introduced a groundbreaking framework that allowed agents to decompose complex tasks into simpler subtasks and dynamically merge learned policies. This work laid the conceptual groundwork for modern approaches to scalable, multi-task learning in robotics and AI, influencing how researchers design systems that can adapt to changing objectives without retraining from scratch. Karlsson’s contributions are particularly notable for bridging the gap between theoretical reinforcement learning and practical robot control, enabling agents to efficiently acquire and combine diverse skills. His research has been widely cited in subsequent studies on hierarchical architectures, transfer learning, and autonomous navigation. For students and researchers, Karlsson’s work remains a touchstone for understanding how intelligent systems can learn to manage multiple, often conflicting goals—a challenge central to advancing autonomous agents in real-world environments.
Research Focus
Key Achievements
Top Papers
- 1