Papers
13
Total Citations
273
H-Index
10
About
Stefan Elfwing is a leading researcher in reinforcement learning (RL) and evolutionary robotics, whose work bridges the gap between biological decision-making and artificial intelligence. His most significant contributions center on developing safe, modular, and biologically inspired RL architectures. Elfwing pioneered the MaxPain algorithm, which parallelizes reward and punishment prediction—a departure from traditional RL that treats punishments as negative rewards—enabling safer autonomous navigation and control in robots. His work on modular deep RL, which decomposes complex tasks into parallel sub-goals, has been highly influential, with his 2020 paper on the topic garnering 60 citations. Elfwing has also advanced embodied evolution, demonstrating how robots can autonomously evolve survival behaviors without human intervention, and developed free-energy based RL methods for high-dimensional state spaces. His research consistently draws inspiration from neuroscience, particularly the separate neural systems for reward and punishment observed in animals. With over 260 total citations across his top ten papers, Elfwing’s work has shaped modern approaches to safe, efficient, and biologically plausible RL systems, making him a key figure in the field.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4Biologically Inspired Embodied Evolution of Survival34 citations · 2005
- 5Darwinian embodied evolution of the learning ability for survival30 citations · 2011
- 6
- 7
- 8Emergence of Polymorphic Mating Strategies in Robot Colonies11 citations · 2014
- 9
- 10