Building Agents from Shared Ontologies Through Apprenticeship Multistrategy Learning

The goal of this research is to create the Disciple Learning Agent Shell for efficient development of personal agents, that relies on importing ontologies from existing repositories using the Open Knowledge Base Connectivity (OKBC) protocol (Chaudhri et al, 1997) and on teaching the agents to perform various tasks through apprenticeship and multistrategy learning (Tecuci, 1998). We are performing significant developments of Disciple, and we are integrating it with OKBC into the following architecture of the Learning Agent Shell: