A Knowledge Request-Broker Architecture for Development of Artificial Social Intelligence

摘要

Artificial social intelligence is about making better intelligent systems out of a group of cooperating less intelligent problem solver agents. Better intelligence can be achieved by having expert agent with diverse expertise cooperating and working on the same problem. For that purpose a platform is required, which, in one hand, provides the freedom for developers to implement any form of solution in their agents, and in the other hand, facilitates mutual understanding between heterogeneous agents by standardizing the way they represent and inquire knowledge. While must published works are concerned with one or the other, the Knowledge Request Broker (KRB) system reported in this paper supports both requirements, and is made specifically for intelligent agents. This paper proposes an architecture for KRB, and provides a practical example for its usage in solving curve-fitting problems.