What happened to expert systems?

I am currently considering using the Expert System for my application.

I really like the basic idea of ​​expert systems, but I have no direct experience with them. While I was searching through the Internet, I realized that expert systems have become significantly less popular than in the past (no new books, outdated software ...).

What is the reason for this decline in popularity? Is his main idea overcome by some new and more effective approaches? Is it worth spending time developing some expert systems, or is it better to try to find a more promising scheme?

I understand that these questions are difficult to answer at all, but I'm just trying to figure out the reasons for this significant breakdown.

Update . There are 10 main categories of expert system applications (interpretation, forecasting, diagnostics, design, planning, monitoring, debugging, repair, training, and control). My main question is the weather . There are several better alternative concepts (for these tasks) that the original expert systems overcome.

I'm mainly interested in the Design category (= setting objects under constraints using a lot of simple expert knowledge).


source to share

1 answer

I studied A.I. in the 70s, and I never really worried about expert systems. The hype seems to have exceeded the content, which seems to boil down to "have a lot of IF-THEN rules and have a system for firing them." To me it looked like an awful concrete about something before understanding it very well.

However, some interesting expert systems have been built. DEC had an expert system for tuning VAX computer systems. Presumably, an expert system is currently being used to plan the airline's gates. I am sure there are many other examples.

If you look at the current online search engines, and IBM Deep Blue (or whatever it is called), it could be called expert systems. The GPS and Siri systems are pretty impressive. At least they seem to know a lot, although they don't seem to like very much to think, understand, reason, learn, etc.

It may not be that much needed for such software these days, compared to more mundane things.



All Articles