Papers
8
Total Citations
609
H-Index
6
About
Thomas Martinetz is a computational neuroscientist and machine learning researcher whose career has been defined by pioneering work at the intersection of neural networks, robotic control, and self-organizing systems. His most influential contributions emerged in the late 1980s and early 1990s, when he developed sophisticated extensions of Kohonen's self-organizing map algorithm to address the challenging problem of visuomotor coordination in robotic systems. His 1989 paper on topology-conserving maps and his 1990 work on three-dimensional neural networks for robot arm control — each accumulating over 260 citations — demonstrated how biologically inspired learning rules, including Widrow-Hoff error correction, could enable robots to learn complex spatial mappings through trial-based experience alone. Martinetz further extended this framework to hierarchical neural architectures capable of coordinating arms with redundant degrees of freedom, reflecting a deep interest in biologically plausible control strategies. His later work explored Hebbian-like adaptation rules and, more recently, manifold-based approaches to adaptive visual sensing, signaling a continued evolution toward efficient, data-driven perception. Across his career, Martinetz has made lasting contributions to the foundations of neural network-based robotics, offering students and researchers a compelling model of how theoretical neuroscience can drive practical advances in autonomous systems.
Research Focus
Key Achievements
Top Papers
- 1Topology-conserving maps for learning visuo-motor-coordination265 citations · 1989
- 2Three-dimensional neural net for learning visuomotor coordination of a robot arm262 citations · 1990
- 3Hierarchical neural net for learning control of a robot's arm and gripper29 citations · 1990
- 4
- 5
- 6
- 7Visual manifold sensing3 citations · 2014
- 8