George D Konidaris

George D Konidaris serves as Associate Professor of Computer Science at Brown University, where his research bridges artificial intelligence, machine learning, and robotics with emphasis on autonomous decision-making systems. His work focuses on developing algorithms that enable robots and AI agents to learn hierarchical structures, discover reusable skills, and operate effectively in complex environments.

Education:

• 2010: PhD, University of Massachusetts, Amherst
• 2003: MS, University of Edinburgh
• 2001: BS, University of the Witwatersrand
• 2000: BS, University of the Witwatersrand

His research spans reinforcement learning, robotic motion planning, and hierarchical abstraction, with significant contributions to skill discovery, temporal abstraction, and model-based methods. Current work integrates visuo-haptic perception for manipulation tasks and explores language-guided robotics using large language models. His approach emphasizes creating systems that learn compact world representations for efficient long-horizon planning in partially observable environments.

Analysis of his 2025 publications reveals strong trends in model-based reinforcement learning with focus on memory mechanisms, uncertainty quantification, and hierarchical skill composition. Key themes include temporal abstraction for planning efficiency, visuo-haptic fusion for robotic manipulation, and language grounding for task specification. His work increasingly connects cognitive science concepts like theory of mind with AI capabilities.

Teaching responsibilities include CSCI 1410 (Artificial Intelligence) and CSCI 2951X (Reintegrating AI), where he bridges theoretical foundations with practical robotics applications.