20
$\begingroup$

The Artificial Intelligence SE is now in private beta phase, and it looks like most questions asked there are on-topic here. There is already another SE with a large overlap, namely Data Science SE. What do you think of this new SE, and should we avoid another almost identical SE?

$\endgroup$
22
  • 6
    $\begingroup$ Smarter users would be a great start. $\endgroup$
    – Sycorax Mod
    Aug 5 '16 at 1:45
  • 1
    $\begingroup$ @GeneralAbrial, ouch. FWIW, I can't see who the users are. $\endgroup$
    – gung - Reinstate Monica
    Aug 5 '16 at 15:14
  • 10
    $\begingroup$ @gung To clarify -- I still don't understand how the Data Science SE got green-lighted when it has about 80% overlap with Stats.SE, and the remaining 20% being composed of stats-specific computing questions. To my mind, AI is either an intractable computational goal (make a computer that can learn arbitrarily complex tasks without human intervention) or is entirely a subset of contemporary machine learning. But either greenlighting decision totally makes sense if one does not really understand what stats.SE does. $\endgroup$
    – Sycorax Mod
    Aug 5 '16 at 15:19
  • 7
    $\begingroup$ It's not up to us to "avoid" it or not. My guess is that what isn't on topic already here, &/or at Data Science, is likely to be on topic on Theoretical Computer Science (eg, they have a tag for ai.artificial-intel w/ 53 questions). I'm not sure if SE (or the larger information-seeking world) is ultimately benefited by increasingly narrow SE sites. I would think many of these would be better combined into a smaller number of stronger sites until there was really critical mass for a split. $\endgroup$
    – gung - Reinstate Monica
    Aug 5 '16 at 15:20
  • 15
    $\begingroup$ Honestly, I think that the strength of SE is that it groups together these things -- so I can contribute to questions about things that I know about and learn about things that I don't know about. Splintering SE sites into narrowly-focused topics makes that more challenging to do: following the goings-on in 3 or 4 places, rather than 1 or 2. I don't see any upside to splintering SE topics. $\endgroup$
    – Sycorax Mod
    Aug 5 '16 at 15:28
  • 2
    $\begingroup$ @gung Perhaps "avoid" isn't the right word, but if a new SE is created and has a 80% overlaps with stats, it might mean some users aren't aware of what is on-topic on stats. On meta.ai.stackexchange.com/q/4/4 I listed different SE that has an AI tag. $\endgroup$
    – Franck Dernoncourt
    Aug 5 '16 at 16:15
  • $\begingroup$ @FranckDernoncourt, that seems like a good thing to do. Note, however, that I cannot see the thread b/c I am not a member there. $\endgroup$
    – gung - Reinstate Monica
    Aug 5 '16 at 16:20
  • $\begingroup$ @gung I think you can join through area51.stackexchange.com/proposals/93481/… $\endgroup$
    – Franck Dernoncourt
    Aug 5 '16 at 16:25
  • $\begingroup$ @FranckDernoncourt, meh. On a different topic, you are probably aware that we have a artificial-intelligence tag. However, it had no excerpt / wiki. I have added one, in case you want to check / tweak it. $\endgroup$
    – gung - Reinstate Monica
    Aug 5 '16 at 16:47
  • 5
    $\begingroup$ I think all the problems around these new websites would can be solved by changing the site name to something like "data science and statistics". CV is clearly too obscure to signal what's on topic. $\endgroup$
    – Sycorax Mod
    Aug 6 '16 at 14:49
  • 1
    $\begingroup$ @GeneralAbrial "CrossValidated" (which I think is a great name) is just a name. As a comparison, the SE site on cooking is called "Seasoned Advice". Its url address is cooking.SE and ours is stats.SE. I recall that SE does not allow site names to differ from their topics anymore, btw. Are you suggesting that we should rather get rid of our "custom name" too? $\endgroup$
    – amoeba
    Aug 6 '16 at 23:35
  • 5
    $\begingroup$ @amoeba I realize that the correspondence between the name and what is on-topic is, essentially, arbitrary. CV.SE could be called "bacon" and we'd still be able to moderate the forums to the same standards. The name is cute and clever for people who understand the field. But the proliferation of increasingly narrow SE sites which are, otherwise, on-topic elsewhere is precisely counterproductive. User kenorb doesn't see any relation between AI and statistics, which is clearly preposterous! Either AI is a farce or statistical concepts are paramount -- I know what side I'm on. $\endgroup$
    – Sycorax Mod
    Aug 7 '16 at 1:34
  • 1
    $\begingroup$ @CliffAB I actually think that statistical coding questions should be on topic but I'm pretty much alone there. Database questions have their own forum. $\endgroup$
    – Sycorax Mod
    Aug 8 '16 at 17:19
  • 3
    $\begingroup$ The problem with the massive overlap between CV & the Data Science forum is that "data science" is to some extent just a buzz word of recent origin, not with CV's catchy name. It's a broad, fuzzy term whose meaning drifts from person to person. The same is true to a lesser extent of the definition of A.I. I'd rather see the D.S. forum closed to avoid such confusion, unless it survives as a mere adjunct of CV strictly for statistics-related coding questions. Until a clear definition of A.I. becomes widespread, I see no point in adding even more confusion with another overlapping site. $\endgroup$
    – SQLServerSteve
    Aug 9 '16 at 11:29
  • 1
    $\begingroup$ @SQLServerSteve, I agree on DS, it's a marketing fad, but A.I. is different. It tries to build a thinking machine. It happens so that lately statistical methods are popular in the field, but that's just a phase in the development of a field. $\endgroup$
    – Aksakal
    Aug 9 '16 at 20:11
9
$\begingroup$

I think that this kind of issue is endemic to SO and will only increase as sites proliferate.

A proposal: there should be a proposal process where users can vote to merge sites, resulting in a permanent redirect.

Really, what's in a name? Cross Validation is an excellent name to catch the attention of anyone who's actually exercised curiosity + healthy skepticism about "smart algorithms" and so on, but inasmuch as it serves as a clue, it's an easy one for people to miss.

$\endgroup$
1

You must log in to answer this question.

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