Description

The goal of this work is to develop knowledge capture interfaces that can reason about the status of the knowledge being developed and provide assistance in building better knowledge that makes the overall problem solving more successful over time. In order to support this, the system maintains a memory model that relates each knowledge to the problem solving episodes it matched, the episodes where it was actualy used, and the episodes it can potentially contribute to. This model is used in assessing user entered knowledge, checking if it can be confidently used, it conflicts with some successful results and needs some modifications, or it needs significant changes including deactivation. Based on this assessment, Echo guides the user in improving and maintaining the confidence levels of the knowledge over time.

More (Research)