Max Robotics — Global Service Robot Directory + US Certification

About

Mark Humphrys is a pioneering researcher in artificial intelligence, with a focus on reinforcement learning, multi-agent systems, and interactive computer games. His most significant contribution is the development of **W-learning**, a self-organising action-selection scheme for systems with multiple parallel goals, such as autonomous mobile robots. Drawing inspiration from Rodney Brooks' subsumption architecture and implementing it through Watkins' Q-learning algorithm, Humphrys created a framework where selfish Q-learners compete to influence an agent's behavior. This work, detailed in papers from 1995 and 2021, has garnered 32 and 7 citations respectively, and remains influential in robotics and AI. In interactive computer games, Humphrys advanced **believability testing and Bayesian imitation**, as seen in his 2006 paper (49 citations), which challenged traditional game AI by integrating strategic planning with motion modeling. His research highlights the practical use of reinforcement learning value functions—not just for decision-making but for quantifying an agent's desires. Humphrys' work bridges theoretical RL with real-world applications, from house robots to commercial gaming, making him a key figure in autonomous systems and human-like AI behavior.