5
$\begingroup$

I just opened the following question:

Are there any learning algorithms with any provable guarantees for manifold learning or manifold regularization?

the question is mainly about algorithms and guarantees that might come with it. The topic is also an intersection of computer science and machine learning. However, the focus is mainly on algorithms. Since manifolds are a topic that is more likely to be known by the (purer) mathematics community of researchers, I was considering migrate this question and put it on MO instead. However, I was not sure if that was a good idea. Obviously, cross-posting is kind of not appropriate (right?). So I was unsure how to proceed. How do people on math overflow feel? Does it belong better here or in cs theory?

$\endgroup$
7
  • $\begingroup$ Also: wait 7 days before migrating. Who knows, maybe you will get an answer there? $\endgroup$
    – Gerald Edgar
    May 4, 2015 at 13:58
  • $\begingroup$ What do you mean by : manifold learning or manifold regularization. $\endgroup$
    – Gil Kalai
    May 4, 2015 at 19:44
  • 1
    $\begingroup$ @GilKalai that is not an easy thing to address in a comment section in a MO forum. I's suggest reading the slides and papers on the following webpage: mit.edu/~9.520/fall14/Classes/manifold.html $\endgroup$
    – Charlie Parker
    May 4, 2015 at 20:18
  • 4
    $\begingroup$ Hi Charlie, you should explain these notions in your question whether on CS theory or on MO overflow. $\endgroup$
    – Gil Kalai
    May 4, 2015 at 21:20
  • $\begingroup$ The e question as stated would only be appropriate in a machine learning forum, not in MO or CS Theory... You have to give much more background. $\endgroup$
    – Dima Pasechnik
    May 9, 2015 at 21:06
  • $\begingroup$ @DimaPasechnik do you mind expanding on what background I need to give? Its not clear to me any background is needed (though I might be wrong). Is the kind of background needed is to explain what manifolds are? Or is the background needed to expand what provable guarantees are? Either should be known by professionals in this field...or am I wrong? Thanks for your help Dima. :) $\endgroup$
    – Charlie Parker
    May 9, 2015 at 23:45
  • $\begingroup$ surely manifolds (smooth? topological?) are well-known to professionals, but manifold learning/regularisation certainly are not. $\endgroup$
    – Dima Pasechnik
    May 10, 2015 at 5:50

1 Answer 1

Reset to default
10
$\begingroup$

I would have nothing against such a question on MO. My guess is that it would be well received here. But I'm not a specialist in the topic, so you should not go by my word only.

If you migrate it from CST to MO, you will lose your CST audience. If you feel that the MO community probably cannot answer all aspects of the question, it might make sense to cross-post. In that case you can tailor the two questions to the different audiences. If you end up doing this, please include links between the versions of the question and explain why you did so. If you want to post on both sites, let the present question be for a while, and only cross-post here if you don't get useful answers for a day or two. Cross-posting will probably not be frowned upon so much if there is a time gap, a link and a rationale.

$\endgroup$

You must log in to answer this question.

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