This paper introduces the concept of learning agent shell as a new class of tools for rapid development of practical end-to-end knowledge-based agents, by domain experts, with limited assistance from knowledge engineers. A learning agent shell consists of a learning and knowledge acquisition engine as well as an inference engine and supports building an agent with a knowledge base consisting of an ontology and a set of problem solving rules. The paper describes a specific learning agent shell and its associated agent building methodology. The process of developing an agent relies on importing ontologies from existing repositories of knowledge, using the Open Knowledge Base Connectivity protocol, and on teaching the agent how to perform various tasks, in a way that resembles how an expert would teach a human apprentice when solving problems in cooperation. This shell and methodology represent a practical integration of knowledge representation, knowledge acquisition, learning and problem solving, and are illustrated with a planning agent that was developed and evaluated as part of the DARPA’s High Performance Knowledge Bases program.
Registration: ISBN 978-0-262-51106-3
Copyright: July 18-22, 1999, Orlando, Florida. Published by The AAAI Press, Menlo Park, California.