Integrating Learning in a Multi-Scale Agent

Ben Weber's dissertation focuses on the goal of building human-level artificial intelligence for real-time strategy games. It explores the following research question: what capabilities are necessary for expert RTS gameplay and how can these capabilities be integrated in a complete game playing agent? The dissertation was produced under the advisorship of Michael Mateas and Arnav Jhala in the Expressive Intelligence Studio at the University of California, Santa Cruz.