Agent assistants for team analysis

With the growing importance of multiagent team-work, tools that can help hu mans analyze, evaluate, and understand team behaviors are also becoming inc reasingly important. To this end, we are creating isaac, a team analyst age nt for post hoc, offline agent-team analysis. ISAAC'S novelty stems from a key design constraint that arises in team analysis: Multiple types of model s of team behavior are necessary to analyze different granularities of team events, including agent actions, interactions, and global performance. The se heterogeneous team models are automatically acquired by machine learning over teams' external behavior traces, where the specific learning techniqu es are tailored to the particular model learned. Additionally, ISAAC uses m ultiple presentation techniques that can aid human understanding of the ana lyses. This article presents ISAAC'S general conceptual framework and its a pplication in the RoboCup soccer domain, where ISAAC was awarded the RoboCu p Scientific Challenge Award.