Mete Tuluhan Akbulut

About

Mete Tuluhan Akbulut is a researcher advancing the frontier of cognitive robotics, with a focus on skill acquisition, learning from demonstration (LfD), and reinforcement learning (RL). His work centers on enabling robots to learn dexterous, adaptive behaviors by combining human-guided demonstrations with autonomous self-improvement. In his highly cited 2020 paper, “ACNMP: Skill Transfer and Task Extrapolation through Learning from Demonstration and Reinforcement Learning via Representation Sharing,” Akbulut proposes a novel LfD+RL framework that allows robots to first imitate a skill and then refine it through exploration—bridging the gap between supervised and trial-and-error learning. This work, with 11 citations, has influenced approaches to efficient robot training. Additionally, his research on affordance learning, published in “Learning Object Affordances from Sensory-Motor Interaction via Bayesian Networks with Auto-Encoder Features,” explores how robots can infer the functional possibilities of objects through interaction, a foundational concept for autonomous manipulation. By integrating Bayesian networks with auto-encoder features, Akbulut contributes to building more intuitive and capable robotic systems. His work is shaping the future of intelligent robotics, where machines learn not just to perform tasks, but to understand and adapt to their environments.

Research Focus

Key Achievements

Top Papers

1. 1
   ACNMP: Skill Transfer and Task Extrapolation through Learning from Demonstration and Reinforcement Learning via Representation Sharing

   11 citations · 2020
2. 2
   Learning Object Affordances from Sensory-Motor Interaction via Bayesian Networks with Auto-Encoder Features

   2 citations · 2020

Key Collaborators

Contact & Links