Papers
6
Total Citations
142
H-Index
5
About
Tanis Mar is a robotics researcher specializing in tool affordance learning, self-supervised machine learning, and humanoid robot cognition, with a particular focus on enabling robots to autonomously understand and use tools. Working extensively with the iCub humanoid robot platform, Mar has made significant contributions to the challenge of teaching robots to generalize tool use beyond their physical limitations — a problem central to the development of truly adaptable autonomous systems. Mar's most influential work, "Self-supervised learning of grasp dependent tool affordances on the iCub Humanoid robot" (2015, 49 citations), demonstrated that robots can learn how different grasps affect a tool's functional utility, a nuanced insight that bridges perception, action, and object geometry. Subsequent research explored 3D geometric representations and parallel Self-Organizing Map (SOM) architectures to predict tool affordances, collectively accumulating over 140 citations across six key publications. Mar's 2017 contributions further advanced the field by showing how 3D tool geometry alone can drive affordance learning without explicit human supervision. Complementing this work, Mar also contributed to depth-driven visual attention systems, strengthening the perceptual foundations needed for reliable robot-environment interaction. Collectively, Mar's research offers a coherent and practically impactful vision of robots that learn, adapt, and act more like humans through experience.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3
- 4
- 5
- 6