Sam Maes

Papers

Total Citations

H-Index

About

Sam Maes is a researcher whose work lies at the intersection of reinforcement learning, probabilistic modeling, and multi-agent systems. His primary contributions focus on tackling the challenge of learning and decision-making in large, complex state spaces—a fundamental hurdle in artificial intelligence. Maes’s most cited work, “Reinforcement Learning in Large State Spaces” (2003, 16 citations), explores scalable approaches to value function approximation and policy optimization, providing foundational insights for modern deep reinforcement learning. In a notable earlier paper, “Q-Learning in Simulated Robotic Soccer” (2002, 3 citations), Maes introduced the innovative use of Bayesian networks to model other agents under conditions of incomplete information, demonstrating how probabilistic graphical models can enhance strategic reasoning in dynamic, adversarial environments. This work bridges reinforcement learning and Bayesian inference, offering a principled framework for handling uncertainty and opponent modeling. Though his citation counts are modest, Maes’s research represents a thoughtful, early exploration of key ideas that would later become central to autonomous systems and multi-agent coordination. His contributions remain relevant for students and researchers interested in the theoretical underpinnings of scalable learning and decision-making under uncertainty.