Jan Drugowitsch

About

Jan Drugowitsch is a researcher whose work bridges computational neuroscience, machine learning, and robotics, with a particular focus on how agents learn and navigate through noisy sensory environments. His key research areas include learning classifier systems, probabilistic filtering, and path integration—the ability to estimate one’s heading from motion cues. Drugowitsch made significant contributions by demonstrating how the XCSF learning classifier system can be enhanced through sensory filtering, improving both learning robustness and robot arm control performance. This work, published in 2013 with 12 citations, showed that exploiting forward velocity knowledge could dramatically refine motor learning. He further advanced the field by providing a probabilistic reformulation of learning classifier systems from a machine learning perspective, offering a rigorous, first-principles foundation for these adaptive algorithms. More recently, Drugowitsch has tackled the challenge of angular path integration, developing a projection filtering method that provides a probabilistic description of heading estimation from noisy velocity observations—a critical capability for both robotic and biological navigation systems. His work is notable for its mathematical rigor and its potential to inform both artificial intelligence and our understanding of neural computation.

Research Focus

Key Achievements

Top Papers

1. 1
   Filtering Sensory Information with XCSF: Improving Learning Robustness and Robot Arm Control Performance

   12 citations · 2013
2. 2
   Filtering sensory information with XCSF

   3 citations · 2012
3. 3
   Learning classifier systems from first principles : a probabilistic reformulation of learning classifier systems from the perspective of machine learning

   2 citations · 2007
4. 4
   Angular Path Integration by Projection Filtering with Increment Observations.

   2 citations · 2021

Key Collaborators

Contact & Links