4
$\begingroup$

Background

There has been a lot talk (both here and CV.SE) regarding if RL questions should be encouraged to be posted here instead of CV.SE or DS.SE. The mods do not seem to agree on which path should be taken. But most of the discussions are quite old; most of them date back to 2016 when AI.SE was still in beta. I am yet to find any thread on the current scope of this forum (suggestions are welcomed).

Question

Is it recommended to questions strongly related to Statistical Learning Theory or Multi-armed Bandits be posted here? Or should CV.SE be our first choice? The issue here is that both historically and by nature they have a strong statistical flavour, although, they find more application in the relatively less stat-heavy field of RL (compared to ML). This discussion, for example, suggests that anything related to the theory of RL should be on-topic here.

$\endgroup$
1
  • $\begingroup$ Yes it is recommended. However I have a different opinion on whether questions should or shouldn't be posted. My idea is that post wherever you are likely to get an answer. The above 2 topics will get you 0 answers on DS.SE and answers only if you post in AI.SE or CV.SE. And we do have users who can answer such questions. $\endgroup$ – user9947 Feb 17 at 10:22
6
$\begingroup$

All theoretical/conceptual questions about statistical learning theory (which is an approach to computational learning theory, developed by Vladimir Vapnik and others, a sub-field of AI) and multi-armed bandit problems/algorithms are on-topic here, but they are also on-topic on Stats SE.

Of course, as a mod of Artificial Intelligence Stack Exchange, I would like to see good/great questions about this topic on AI SE. So, of course, I will suggest that you first ask this type of question here. However, after a while, if you don't get an answer, you probably should ask it somewhere else, such as at Stats SE.

In general, the choice of where to ask your question is yours and will probably depend on

  • whether it's on-topic or not (it's in both cases, so, in this case, it should not depend on this),
  • the service that you received in the past was good or not (did you get a good answer there when you asked a question there?)
  • your preferences (do you like more AI SE/Artificial Intelligence or Stats SE/Statistics?)
$\endgroup$

You must log in to answer this question.

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