Rate My Professor Daniel Polani

Professor Daniel Polani serves as Professor of Artificial Intelligence in the Department of Computer Science, School of Physics, Engineering & Computer Science at the University of Hertfordshire. He is a key member of the Centre for AI and Robotics Research and coordinates the Adaptive Systems research group. Polani's academic career at the university spans research activity from 1992 to 2025, with 185 research outputs documented. His primary research interests center on artificial intelligence, particularly information-theoretic methods for modeling cognition, perception-action loops, and adaptive behavior in complex environments. He has developed foundational concepts such as empowerment, a universal utility function measuring an agent's potential for environmental influence through channel capacity in the sensorimotor loop, promoting goal-agnostic intrinsic motivation in robotics and reinforcement learning. This approach addresses challenges in high-dimensional state spaces and distributed systems, such as 3D-printed robots with limited processing capacity.

Polani's influential publications include 'Information flows in causal networks' with N. Ay (Advances in Complex Systems, 2008), cited over 400 times; 'Empowerment for Continuous Agent-Environment Systems' with T. Jung and P. Stone (2012); and recent contributions like 'Process empowerment for robust intrinsic motivation' with S. Tiomkin and C. Salge (Journal of Physics: Complexity, 2025), 'SuPLE: Robot Learning with Lyapunov Rewards' with P. Nguyen and S. Tiomkin (ICRA 2025), and 'Dimensionality Reduction of Dynamics on Lie Groups via Structure-Aware Canonical Correlation Analysis' with W. Chung and S. Tiomkin (ACC 2024). His work has garnered over 6,200 citations, an h-index of 39, and an i10-index of 105 according to Google Scholar. As principal investigator, he has secured funding for projects including CORBYS (2010-2015), WiMUST (2015-2018), and socSMCs (2015-2019) under EU Horizon 2020, alongside ongoing initiatives like Information Constrained Control funded by The Pazi Foundation. These efforts advance cognitive robotics, marine robotics, and self-organization in multi-agent systems.