Logic Programming Thread

I tried it once and have no interest in reliving that god awful experience ever again.

>>59297948
I just learned it in Uni. I fucking hated it at times, but ended up really enjoying it. Such a different way of thinking. It also really gave me a much better intuition for recursion.

Don't know enough about if it'll make a comeback since its big thing was AI and linguistics, and the former is all about neural nets now. Don't know if linguists still care about it, either.

I recently looked up real-world Prolog applications and there were a surprising amount for NoSQL databases.

Related, found this recently:
https://linux.slashdot.org/story/15/06/30/0058243/interviews-linus-torvalds-answers-your-question

>We'll get AI, and it will almost certainly be through something very much like recurrent neural networks. And the thing is, since that kind of AI will need training, it won't be "reliable" in the traditional computer sense. It's not the old rule-based prolog days, when people thought they'd *understand* what the actual decisions were in an AI.

>>59298143
some quotes Ive collected from plebbit

AI is huge. Neural networks is not the only kind of AI. In the old fashion, there are at least three kinds of AI, such as Symbolism, Connectionism and Behaviorism. This aspect of classification may be obsolete because modern algorithms usually have those two or all property.
Symbolism is come from the oldest formal inference system called logic, Connectionism is come from biological neural networks, and Behaviorism is actually a methodology of psychology.

Currently popular approaches include statistical methods, computational intelligence and traditional symbolic AI.

Why exactly is Lisp considered an AI programming language?
While a lot of replies have given a correct historical answer to your exact question, (namely, "Because that's where Lisp comes from"), I want to answer the closely related question, "Why is Lisp good for AI?"
Firstly, while we can argue that Lisp is good for Good Old Fashioned AI (GOFAI), which is the logically-grounded symbolic approach to AI that McCarthy and his followers had in mind, it's quite possible that if you want to do non-symbolic AI, such as neural networks, support vector machines or other kinds of statistical machine learning, you're better off with a fast, statically-typed language with good numerical performance like C or C++. If your system uses both symbolic and non-symbolic techniques, the decision between Lisp and a static language is harder.

ML is a very different field from the old school AI that made heavy use of lisp.
Old AI was all about symbolic manipulation, and for this lisp is king.
ML fundamentally relies upon vectors, matrices, and tensors. For this some kind of array-based programming language (I recommend numpy and python) is much more important.
I did try using common lisp for ML a long time ago. I ended up implementing in macros array-broadcasting, and with hindsight it was remarkably close to a shitty version of numpy.