Knowledge granularity spectrum, action pyramid, and the scaling problem

In this paper we introduce the concept of knowledge granularity and study t he relationship between different knowledge representation schemes and the scaling problem. By scale to a task, we mean that an agent's planning syste m and knowledge representation scheme are able to generate the range of beh aviors required by the task in a timely fashion. Action selection is critic al to an agent performing a task in a dynamic, unpredictable environment. K nowledge representation is central to the agent's action selection process. It is important to study how an agent should adapt its methods of represen tation such that its performance can scale to different task requirements. Here we study the following issues. One is the knowledge granularity proble m: to what detail should an agent represent a certain kind of knowledge if a single granularity of representation is to be used. Another is the repres entation scheme problem: to scale to a given task, should an agent represen t its knowledge using a single granularity or a set of hierarchical granula rities.