Inferring relations and individuals relevant to a situation: An example

摘要

In various scenarios, agents need to exchange information about situations — either the situations they are currently in or in reference to other cir-cumstances. Assuming that agents are intelligent agents, i.e., they have their own inference capabilities, the question then is — which parts of the knowledge that an agent has need to be sent over (possibly narrow bandwidth) communication links in order to provide a complete description of the situation? Or in other words, which part of the agent's knowledge is relevant to a specific situation? The assumption behind this question is that only a small part of the agent's knowledge needs to be sent since the receiving agent can infer the rest of it locally. This paper discusses an example of reasoning about the relevancy of information. First, the notion of relevancy is defined. Then an implementation of relevance inference scenario is described.