SHERPA:

Guidance for Knowledge Acquisition in Large-Scale Knowledge Bases

This is a summary of the participation of the EXPECT project at USC/ISI in DARPA's new program on High-Performance Knowledge Bases .

This work will build on the EXPECT architecture and extend it in three ways:

EXCALIBUR: the next generation of the current EXPECT system, our framework for developing KBSs that includes knowledge acquisition tools. EXCALIBUR will include an extended language for problem-solving knowledge and an extended KBS compiler.
SHERPA: an acquisition mediator that will support knowledge acquisition for KBSs not developed with our framework. The mediator will have (possibly incomplete) models of the KBS reasoning, and will use them to selectively route updates to the knowledge library.
LIBRA: a framework for representing and managing a library of problem-solving methods, including indexing tools, composition tools, and selection tools.

A slide show is available from the presentation at the West Coast kickoff meeting that took place in March, 1997.

Relevant parts of the technical proposal, entitled "Knowledge Acquisition for Large Knowledge Bases: Integrating Problem-Solving Methods and Ontologies into Applications" can be found here .

For more information about EXPECT, send mail to expect@isi.edu.

Last updated: March, 1997.