Mixed-Initiative Knowledge Base Development

Abstract

This research has developed an end-to-end mixed-initiative approach to the development of knowledge bases by subject matter experts, with limited assistance from knowledge engineers. In this approach, the complex knowledge engineering activities, traditionally performed by a knowledge engineer and a subject matter expert, are replaced with equivalent ones performed by the subject matter expert and a learning agent, through mixed-initiative reasoning, with limited assistance from the knowledge engineer. In essence, the learning agent helps the subject matter expert to describe a specific problem, to make explicit how he or she solves it, to formalize this reasoning and to explain it to the agent. At the same time, the expert helps the agent to understand this reasoning process, to learn general problem solving tasks and rules from it, and to refine its ontology, thus developing its knowledge base to represent the expertise of the subject matter expert. These methods have been implemented in the Disciple-RKF learning agent that has being successfully used in several courses at the US Army War College, in the context of the center of gravity analysis problem. Experimental results demonstrate that the developed methods simplify the acquisition of knowledge and improve the knowledge base development process.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Aug 31, 2003
Accession Number
ADA417711

Entities

People

  • Gheorghe Tecuci

Organizations

  • George Mason University

Tags

Communities of Interest

  • Autonomy

DTIC Thesaurus Topics

  • Acquisition
  • Artificial Intelligence
  • Center Of Gravity
  • Engineering
  • Engineers
  • Gravity
  • Information Systems
  • Intelligent Agents
  • Intelligent Systems
  • Machine Learning
  • Military Applications
  • Military Operations
  • Ontologies
  • Operations Research
  • Software Development
  • Students
  • War Colleges

Fields of Study

  • Computer science

Readers

  • Artificial Intelligence