Lambda the Ultimate

It's interesting to investigate Doug Lenat's track record. He produced a number of very interesting AI applications in the 70s, including AM (Automated Mathematician) and Eurisko. There have been several write-ups on these programs and they seem mind-blowing. [snip]

Hehe, a combinatorial reasoning system thought was pestering the back of my mind so I guess it's no wonder I ended up at his work ;-)

And then in the 90s Lenat moved onto Cyc. For well over a decade there were promises that it was going to change everything. Every time the promises were a little toned down from before. I worked briefly with someone who had actually worked for Cycorp and he believed that basically there was "nothing there".

To some extend I guess capitalizing on your own ideas is the ultimate test; and a fast manner of developing a mid-life crisis ;-)

You can dig up papers on AM and Eurisko and it all looks vaguely plausible. But then in the early 80s a lot of things looked plausible that today don't.

Whitehead said it better than I can:

Systems, scientific and philosophic, come and go. Each method of limited understanding is at length exhausted. In its prime each system is a triumphant success: in its decay it is an obstructive nuisance.

I really have no idea what to make of Doug Lenat from all of the second hand opinions I've heard. But I'm pretty sure there's a book with a great story that needs to be written about him and his work.

Second hand opinions? Yeah, audiences and the art of the narrative. The average John Doe doesn't really want to understand e = mc² but what affairs Einstein had. Fortunately, most scientists take being dull as an artform so let's stick to discussing ideas.

His thing seems to be based on experiments with ontological structures and algorithms, can't see anything wrong with that so I guess you're right and his ideas deserve an entry somewhere.

Unfortunately everything I know about AM and Eurisko was gleaned from accounts published on paper so apologies for the lack of links.

In some respect the Internet and Google is the largest expanding interacting ontological structure we have, so that really is no problem. Thanks for the information.

BTW: thought I add some cut and paste stuff on Eurisko because the link seems unstable [www.daxtron.com]:

EURISKO by Doug Lenat is a program that showed impressive performance in numerous domains using "learning by discovery." EURISKO not only learned new domain-specific definitions of concepts but also new heuristics that guided its search for new concepts. EURISKO learns about solving specific problems and also about learning in general. EURISKO has access to information about itself and modifies itself to perform better over time. EURISKO explores possible configurations of concepts and ranks them as to "interestingness." The program was the champion of a naval war game tournament for two years, designed several innovative 3-D very large-scale integrated circuits, and was applied successfully to several other domains.

EURISKO showed it was possible for Artificial Intelligence systems to:

• Make significant contributions to little-explored domains
• Automate discovery by using simulation
• Use heuristic search for domains too large to explore manually
• Be applicable to domains where no human experts yet exist
• Create novel plans and system capable of beating numerous adversaries

EURISKO relied on a representation where syntax mirrors semantics, and therefore achieved a great deal by syntactic exploration and mutation. However, not being limited by common sense ultimately became its biggest weakness. Other problems with EURISKO included its large size, long running time, and its lack of specific direction. EURISKO no longer exists in usable form. The early success of Eurisko helped inspire the development of both genetic programming and CEMA.