Applied AI research with StarCraft, Torch and the BEAM

Barbara and Jean will present their community bot system built for online competition and AI research on the real-time strategy game of StarCraft. This bot system is a message-passing, asynchronous system that exploits the hot swap loading, parallelism and standard mechanisms of Luerl and the properties of the BEAM ecosystem.

OBJECTIVES

Barbara and Jean will present their integration between StarCraft and the BEAM as an interesting domain for applied AI research, since represent a well defined adversarial environment which pose a number of interesting challenges in areas of information gathering, planning, dealing with uncertainty, domain knowledge exploitation, task decomposition, spatial reasoning and machine learning.

TARGET AUDIENCE

Machines of all ages, curious developers, software engineers, machine learning enthusiasts, gamers and non-gamers, people in general who want to know about the guts and implementation of the presented work in the StarCraft AI competition arena, with a polyglot system that's about building bridges between communities and language barriers.