11
$\begingroup$

What is our stance on questions about non-statistical AI? Are they on-topic or off-topic?

I ask the question to know whether https://ai.stackexchange.com/ has a partial or perfect overlap with CV.

Can you please provide a couple of short examples of such questions? – amoeba ↵ 5 mins ago

I saw some tags that seem to indicate non-stat AI may be on-topic, e.g.:

Examples of questions:

$\endgroup$
6
  • $\begingroup$ Can you please provide a couple of short examples of such questions? $\endgroup$
    – amoeba
    Aug 10, 2016 at 17:02
  • $\begingroup$ @amoeba added~~ $\endgroup$
    – Franck Dernoncourt
    Aug 10, 2016 at 17:10
  • 1
    $\begingroup$ I am somewhat confused by this question. You give three examples but all of them are not closed (and some are highly upvoted and have a lot of answers from high rep users). I conclude that they on-topic. What makes you think that they might be off-topic? (Myself, I have no idea what "non-statistical artificial intelligence" is, so I am simply judging by your three examples.) [Initially, I was expecting you to copy-paste some Qs from AI.SE in order to discuss if they would be on topic here.] $\endgroup$
    – amoeba
    Aug 11, 2016 at 21:13
  • 2
    $\begingroup$ @amoeba sometimes rules aren't applied. $\endgroup$
    – Franck Dernoncourt
    Aug 11, 2016 at 21:17
  • $\begingroup$ The example questions linked in the OP sound pretty statistical to me,. For example ai.stackexchange.com/questions/15 ai.stackexchange.com/questions/1348 ai.stackexchange.com/questions/1768 are more clearly non-statistical AI questions (and off-topic here, I would say) $\endgroup$
    – Juho Kokkala
    Aug 30, 2016 at 12:51
  • 2
    $\begingroup$ Fuzzy logic is a different way to mathematical describe uncertainty. But it is different from probability. Does that make it off topic? $\endgroup$
    – Michael R. Chernick
    Dec 31, 2016 at 22:41

2 Answers 2

Reset to default
5
$\begingroup$

As a mental model, it can be helpful to consider the scope of Stats.SE and AI.SE as two circles of a Venn diagram.

  • Some topics, like or statistical hypothesis tests, are only on-topic on Stats.SE.
  • Some topics, like and , are on-topic on both sites.
  • Some topics, like philosophical or social consequences of AI, are only on-topic on AI.SE.

(And, of course, there are lots of topics that aren't on-topic on either site but are on-topic somewhere else, or no where else, on Stack Exchange.)

Therefore, it's best to ask questions about non-statistical AI on AI.SE. If any user is unsure where to post a question, you can ask about it on stats.meta.SE and we'll point you in the right direction.

$\endgroup$
2
$\begingroup$

There are topics, like

  • expert systems (and, more generally, symbolic AI),
  • artificial general intelligence,
  • meta-heuristics
  • evolutionary computation (swarm intelligence, evolutionary algorithms, etc.),
  • traditional search algorithms (such as A*),
  • philosophical and social aspects of AI (such as super-intelligence, tests and definitions of intelligence, etc.)

that, as far as I know and interpret the on-topic page here, are not on-topic here, but are on-topic on AI SE. Some of them have little to do with statistics or machine learning (though some may disagree, as there may be a way to study these from a statistical point of view). So, there is not a perfect overlap between Artificial Intelligence SE and Cross Validated SE. However, there's a big overlap between the two sites, especially because machine learning (a big part of the two) is on-topic on both sites. Nevertheless, AI SE and Stats SE are significantly different to be separate sites. In any case, we should probably collaborate more, in some way, when it comes to machine learning questions. For instance, we could "advertise" our ML questions to each other, or something like that, so that askers can get an answer more quickly or likely.

The site that I really think has an almost perfect overlap with Stats SE is Data Science SE, though some may disagree (because Data Science SE maybe covers certain topics that are not very welcomed here, such as certain implementation questions).

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .