What languages ​​are used in AI research today?

Question

I am currently involved in expert systems, emacs lisp, and reading about artificial intelligence. Traditionally, artificial intelligence is associated with LISP and expert systems with CLIPS. However, I noticed in computational science how much Python is used. What about the field of artificial intelligence and machine learning? is LISP still dominant? how much python is used in AI? are any of the new functional languages, for example, clojure, used in research?

in the field of expert systems, which shells are most commonly used / popular today? are there any interesting events that you know about the language?

Answer 1

How about artificial intelligence and machine learning?

It depends on the area of ​​AI.

is LISP still dominant?

Not at all, unless it has resumed in the last few years.

how much python is used in AI? are any of the new functional languages ​​clojure for example, being used in research?

People will use any language appropriate for their domain. You should also consider the availability of libraries - if you are performing any data mining task, you can download Weka and use Java. Or, if you are researching agent theory that involves sending structured XML messages, you may find that you are using some kind of language that your favorite XML parser is in. If you are creating a highly efficient backgammon program, you can stick with C / C ++ and use the existing evaluation function (as in my research using gnubackgammon).

in the field of expert systems, which shells are most often used / popular today? is there an interesting language event that you know about?

Expert systems are actually not so much rage. As studies progressed, they were found to be fragile systems that needed too much manual TLC. If you try to detect cancer or qualify someone for a credit card, you are better off using some ML system, such as the Bayesian Network.

I think you will find that most AI textbooks are rather agonistic for the platform, since the basic algorithms are agnostic for the platform.

Answer 2

DISCLAIMER: I'm just an enthusiast, and you can find out even more about AI at the moment, but my research has told me that ...

Python rules in MIT - so the work done there is done in Python.
Prolog, Erlang, Scala, Java, Ruby: all of them also appear on the Internet.
C / C ++: because you always need to do something with a lot of crunches really really fast.

Answer 3

The reason Lisp is connected to AI is because when AI was really new, there weren’t so many of them, and the other languages ​​of the day were unacceptable. Now there is much more diversity in the AI ​​field, and most languages, at least until the 70s, are LISP. Thus, there is not a single dominant language.

If you are interested in AI, stop worrying about the language and start teaching methods. Language does not matter.

---

Other notes:

I study computational linguistics (which is an area of ​​AI that often uses machine learning methods) at Indiana University, and Python is the most popular choice for prototyping in IU. But not the vast majority. In general, computational linguistics does not have a dominant language for research. A lot of software has been released in Java and C / C ++ because these languages ​​are fast and portable. (I run a couple of Java parsers in the background when I enter this value.)

I only know one or two groups that still use Common LISP. I have not heard anyone do research using Clojure, but it is very good there. As for other new functional languages, F # is used by Microsoft to develop AI-esque things, but not necessarily for research. I personally use Haskell for many things, but this is not a general choice as far as I know.