Doug Lenat

"The chief bottleneck in building large AI programs, such as expert systems, is recognized as being knowledge acquisition. There are two major problems to tackle: (1) building tools to facilitate the man-machine interface, and (2) finding ways to dynamically devise an appropriate representation. Much work has focused on the former of these, but our experience with AM and Eurisko indicates that the latter is just as serious a contributor to the bottleneck, especially in building theory formation systems. Thus, our current research is to get Eurisko to automatically extend its vocabulary of slots, to maintain the naturalness of its representation as new (sub)domains are uncovered and explored."

Why AM and Eurisko Appear to Work.

January 1983 Source: DBLP Conference: Proceedings of the National Conference on Artificial Intelligence. Washington, D.C., August 22-26, 1983.

Authors:

Comments

Popular posts from this blog

Perplexity

Aphorisms: AI

DeepAI's Austen on China