Criticism-Based Knowledge Acquisition for Document Generation

All Papers

This paper describes an ai application called the TRADOC issue management expert (time), which was deployed in 1990 for the U.S. Army Training and Doctrine Command (TRADOC). time helps generate a document that is part of the system acquisition milestone or decision process. The goal of time is to bridge the gap between the headquarters decision makers who know how to write the desired document and the various authors at 17 sites nationwide who know what to put in the document. time has several capabilities. In one capacity, it is an expert critiquing and tutoring system that emulates the headquarters decision makers. In this capacity, it communicates good document-authoring rules, heuristics, and practices in real-time (during authoring) mode. In another capacity, time is also a knowledge-acquisition system that interviews the author and helps him(her) generate the document. Finally, time acts as an intelligent memory that dynamically indexes and collects as many as 600 documents that are produced annually in a corporate memory for later analogical retrieval and illustration purposes. This last capability means that the more time is used, the more useful it is.