Fabrizio Smeraldi
École Polytechnique Fédérale de Lausanne, Queen Mary University of London
Papers
5
Total Citations
148
H-Index
5
About
Fabrizio Smeraldi is a researcher whose work spans computational neuroscience, autonomous robotics, and computer vision, with a particular focus on biologically inspired approaches to spatial cognition and machine learning. His most influential contribution — a 2004 paper garnering 96 citations — demonstrated how autonomous agents could achieve spatial learning and navigation through nonuniform Gabor space sampling combined with unsupervised growing networks and reinforcement learning, offering a powerful framework for representing continuous spatial environments from high-dimensional sensory input. Building on earlier foundational work modeling hippocampal place cells and head-direction cells in miniature robots, Smeraldi showed how visual and proprioceptive stimuli could be integrated via Gabor filters and Hebbian learning to replicate key features of mammalian spatial navigation. His subsequent research extended these ideas into humanoid robotics, developing online multiple instance learning algorithms for hand detection under minimal supervision. More recently, his work on the Partial Order Kernel addresses the demanding challenge of visual localization across seasonal and lighting changes — a critical problem in real-world robot deployment. Across his career, Smeraldi has consistently bridged neuroscientific principles and practical machine learning to advance intelligent, perception-driven autonomous systems.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4
- 5