Monkeys, Queens and Cats

Greetings:

[Short rant]

I have been doing AI type work since the late 80's (seems like only yesterday) and, like most who start in AI work or any kind of programming, we start with very simple problems. The first is always the ubiquitous "Hello World" program. The second, if you are learning inheritance or something of that nature, is checking to see if a child is a grandchild of another person or object or checking to see if this object or person is human or a toad and therefore has five fingers on each hand. Something like that.

For most who use rulebase tools we learn (or were supposed to have learned) the problem of how to help direct a monkey to find bananas to satisfy its hunger OR we learn how to solve the constraint problem of four, eight or sixteen queens on a chess board so that they none can capture the other one. The other one (thanks to the Monday night CBS TV series of "The Big Bang") is Schrodinger's cats. I really do think that these are excellent problems for initial learning experience in our little world of AI or its subset of rulebased systems.

Also, I consider MAB the best teaching tools for AI that we have - but it is, after all, a teaching tool for AI geeks, not something that the normal person outside of AI would ever consider. My Daddy used to say that people who do things like that are Educated Idiots or just Book Smart - meaning that they could solve complex problems in class but couldn't hammer a nail into a board nor saw a straight line and, therefore, useless in the real world.

The objection with continuing to solve and/or teach these problems is that once we have done them in class we tend to use them over and over and over in more advanced discussions as though we could not think beyond the beginner level of thought. Whenever we have to explain something outside of our close circle of friends and workmates, we always run back to our "comfort zone" and show "Hello World" or MAB. Also, to those outside of AI, who have never had to solve anything more complex than balancing their checkbooks, these problems seem rather silly and beyond the ken of the ordinary mortal and therefore not something for every day life.

PLEASE! Let's try and bring something to the table that (1) the ordinary mortal can understand and (2) that possibly has some real-world application(s). For example, what if we (a) had a problem in forecasting, or (b) a problem of finding a terrorist in L.A. or (c) how to process fluid flow in a petro-chemical plant so that we ensure that it will not blow up [and what to do if there is that possibility] or (d) complex shift scheduling or maintenance crew assignment or (e) complex resource assignment in a large manufacturing plant or (f) how to configure specialty helicopters?

OK, these are just thoughts. But if we, the AI community, can begin to establish some better examples then maybe we can be accepted and assimilated into the mainstream river of human thought. Maybe...

SDG

jco

Rulebase Benchmarks 2009 Part 3

Greetings:

Once again we return to the world of Rulebase Benchmarks. Benchmarks for rulebased systems seem to have three camps: Those who love them and usually have good numbers to report. Those whose benchmarks are not remarkable one way or the other and are not worried about them; yet. Those who benchmarks are not very good and keep telling their users and/or customers, "Yes. That's OK for academics. But those benchmarks don't represent the 'real world' in which we have to live and work every day."

At October Rules Fest 2009 I asked the T3 (Thursday Think Tank) if they thought benchmarks were important. Dr. Forgy (who has the fastest rulebase engine on the planet Earth - called TECH) and myself (who has spent the past decade exploring benchmarks, their cause and effect) were ardently in favor of such things. FICO (Blaze Advisor) folks were kind of OK with the idea, especially since they use Dr. Forgy's engine in their rulebase. The rest of the group, whose performance was anywhere from just barely OK to abysmal, were not really against using benchmarks but it just wasn't something that they wanted to discuss in public.

So, here's my proposal: If anyone can produce a "real world" benchmark that can be run on any rulebase engine on any platform, PLEASE DO SO!! Otherwise, those of us who are in this for the long haul will continue to examine a rulebase from many different viewpoints but we shall NOT throw out tried and true tests that continually disclose fatal faults in fancy products that will not perform with large data sets combined with large rule sets. Such applications are large banks, large insurance companies, Homeland Security along with "real world" AI problems in rulebased forecasting, petro-chemical processing problems, electrical power grid production problems, shift scheduling for large plants and/or hospitals and anywhere else that will have the many objects - many patterns complex cross products.

Speaking of which, I guess it's about time to publish the 2009 Benchmarks. Our goal is to finish up in December some time so that at least we can say that we finished this year with a completed product. :-)

SDG

jco