Memory-Based Expert Systems

Abstract

There are 3 major areas of accomplishments in recent AFOSR sponsored AI research at Yale. While case-based reasoning is simple in concept, there are of course many subtle and difficult design issues that have to be resolved to make it work. One major achievement in the past two years at Yale has been the development of an explicit structure for the case-based reasoning process, as shown in the slide containing the flowchart entitled 'Case-Based Reasoning'. This process description summarizes results from the JUDGE (Bain 1986) AFOSR project as well as from other projects such as CHEF (Hammond 1986) and COACH (collins forthcoming). The flow of control in the process (represented by the arrows and boxes) is task-independent, while the particular knowledge bases (represented by the ovals) are task specific. The JUDGE program models the subjective assessment task of judicial sentencing. That is, given a description of an event, such as a fight between two people that ended in a death or serious injury, where someone has been convicted of a criminal act.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jun 01, 1987
Accession Number
ADA190203

Entities

People

  • R. Schank

Organizations

  • Yale University

Tags

Communities of Interest

  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Accuracy
  • Artificial Intelligence
  • Cognitive Science
  • Computer Programming
  • Computer Science
  • Construction
  • Criminals
  • Expert Systems
  • Explosions
  • Feedback
  • Information Processing
  • Language
  • Plastic Explosives
  • Reasoning
  • Standards
  • Street Drugs
  • Thinking

Readers

  • Criminal Law
  • Neural Network Machine Learning.
  • Theoretical Analysis.