Deep Reinforcement Learning for Games

With over three billion players in the world, AI is poised to transform the landscape of gaming experiences and the games industry itself. Microsoft’s vision for gaming is a world where players are empowered to play the games they want, with the people they want, whenever they want, where-ever they are, and on any device. In close collaboration with the Xbox Gaming division, we drive towards this transformation through world-leading deep reinforcement learning research.

Projects

Project Malmo

Our research on multi-agent learning aims to develop intelligent agents that can collaborate with people, in applications ranging from video games to assistive technology. As we endeavour to unravel the principles of multi-agent learning and collaboration, our research is facilitated by the Project Malmo, our open-source experimentation platform built on the game Minecraft.

Project Paidia

The focus of Project Paidia is to drive state of the art research in reinforcement learning to enable novel applications in modern video games, in particular: agents that learn to collaborate with human players.