Two Problems in Case-Based Reasoning.
Abstract
During the three years of this fellowship students have studied the use of experience in problem solving. While problem solving is a much-studied area within Artificial Intelligence, traditional research in problem solving have not considered how repetition can be used to advantage by a problem solver. In the research conducted under this fellowship, the authors have been studying how a problem solver can take advantage of repetition and can use its experience to better its performance. The processes that capitalize on repetition are called case-based reasoning processes. Through case-based reasoning, a system can make use of reasoning shortcuts, can avoid previously-made mistakes, and can be alerted to the potential for problems. Work on this project has been mostly in the areas of plan modification and integrating memory and problem solving processes. A program, called the Consumer Advisor, or CAS, Implements the ideas. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Jan 06, 1987
- Accession Number
- ADA176512
Entities
People
- Janet L. Kolodner
- Roy Turner
Organizations
- Georgia Tech