Live data from Hacker News

The Power of Prolog

metalevel.at

161–162 of 162 posts

Re: The Power of Prolog

#161

Prolog was one of hot A.I. prospects of the 1980s. It could be a way to build an expert system. Japanese fifth generation computers were supposed to,run it fast. Then A.I. boom fizzled. Neither the first or last business cycle.

The fifth generation project didn't actually use prolog, but instead a custom prolog-like language.

However, there's still work being done adjacent to 5G using real prolog, by people who thought it was an interesting idea: https://github.com/bitlaw-jp/the-constitution-of-japan

Prolog is still used in the development of expert systems, though when those systems are shipped, they have often been rewritten in an imperative language based on a Prolog prototype.

The fifth generation project, from what I understand, died because of a dedication to custom hardware using already-obsolete fabrication tech -- mid-70s flipflops-on-IC tech over late-70s microprocessor-on-IC tech, wire wrap, etc. There was an inability or unwillingness to iterate on original committee designs. By the time it folded in 1998, everything involved looked like stone knives and bearskins. Those boxes wouldn't have been able to run it fast by 1980 standards.

Re: The Power of Prolog

#162

I imagine that there must be some system similar to prolog but with a probabilistic model of inference, for example ground rules could be sorted by statistical properties to accelerate inference. Just like machine learning trees select features for classification (perhaps using cross-entropy minimization). Edit: I found Journal of Machine Learning Research 7 (2006) 307–342, Kernels on Prolog Proof Trees: Statistical…

I wrote something like this, based in part on ideas from PLN: https://hackernoon.com/mycroft-a-predicate-logic-language-ov...
Post reply on HN