On Behavior Acquisition and Classification in Adversarial Environments

ABSTRACT Increasingly in complex domains with multiple intelligent agents, each agent must be able to identify what the other agents are doing. This is especially important when there are adversarial agents inferring with the accomplishment of goals. Once strategies are identi ed, the agents can then respond to recent strategies and adapt to improve performance. We present an approach to doing adaptation which relies on classi cation of the current adversary into predened adversary classes. For feature extraction, we present a windowing technique to abstract useful but not overly complicated features. The feature extraction and classi cation steps are fully implemented in the domain of simulated robotic soccer, and experimental results are presented.