Multi-agent systems are an attractive approach to modeling systems of interacting entities, but in some cases mathematical models of these systems can offer complementary benefits...
Abstract. The Virtual Reality Modeling Language (VRML) and Java provide a standardized, portable and platformindependent way to render dynamic, interactive 3D scenes across the Int...
The work proposes a hierarchical architecture for learning amic scenes at various levels of knowledge abstraction. The raw visual information is processed at different stages to g...
Controlling the sensing of an environment by an agent has been accepted as necessary for effective operation within most practical domains. Usually, however, agents operate in par...
In this paper we present RETALIATE, an online reinforcement learning algorithm for developing winning policies in team firstperson shooter games. RETALIATE has three crucial chara...