Guner Orhan

Papers

Total Citations

H-Index

About

Guner Orhan is a researcher in cognitive robotics, whose work focuses on endowing humanoid robots with the ability to learn and use context, concepts, and language through sensorimotor interaction. His major contributions center on developing probabilistic frameworks for grounded concept learning, where robots build internal models of their environment from direct physical experience. Orhan pioneered the use of Latent Dirichlet Allocation—a technique borrowed from computational linguistics—to model context as a latent variable in robotic systems, allowing robots to understand and adapt to different situations. He also introduced the "Probabilistic Concept Web," inspired by the human concept web hypothesis, which represents concepts and their relationships using Markov Random Fields. This work has been cited over 50 times, with his most influential paper, "Learning Context on a Humanoid Robot using Incremental Latent Dirichlet Allocation," receiving 20 citations. Notably, Orhan also explored the co-learning of nouns and adjectives, demonstrating how robots can simultaneously acquire object properties and categories in an integrated manner. His research bridges machine learning, cognitive science, and robotics, offering a path toward more autonomous, context-aware artificial agents.