A Mixed-Initiative Approach to Rule Refinement for Knowledge-Based Agents

Abstract

This paper presents a mixed-initiative approach to rule refinement in which a subject matter expert collaborates with a learning agent to refine the agent’s knowledge base. This approach is implemented in the Disciple learning agent shell and has been evaluated in several agent training experiments performed by subject matter experts at the US Army War… (More)

5 Figures and Tables

Topics

  • Presentations referencing similar topics