From the description of cross validate and computational science it seems that cross validate is the correct site to ask question about Machine learning? Am I right to say that? Or what site is better for asking question about machine learning?

  • $\begingroup$ I will say, I think cross validated is supposed to be more tailored to statistical/machine learning type questions. You could surely ask here, but I could see either place being good. One thing to keep in mind is questions get asked frequently on Cross Validated, so it can be tough to ensure you get an answer since your question might disappear off the front page. $\endgroup$
    – spektr
    Jul 8 '16 at 4:11

In principle, a question about machine learning would not be off-topic on either site. I'd break it down further about what kind of question about machine learning you have:

  1. If it's about the application of machine learning to problems in science and engineering, ask here.

  2. If it's about the application of machine learning in data mining or similar, ask at Cross Validated.

  3. If it's about specific machine learning algorithms, their behavior or implementation, ask at Cross Validated (but for the latter, make sure it's on topic there).

  4. If it's about the statistical theory behind machine learning, ask at Cross Validated.

  5. If it's about the mathematical theory, in particular of the optimization methods used in machine learning, ask here.

In short, you should decide which kind of expertise you need to answer your question: Here, you'll find mainly (applied) mathematicians and (computational) physicists, chemists, biologists etc., while Cross Validated has mainly statisticians (although there is some overlap).

One thing to keep in mind is that Cross Validated is a much larger site; this means that there are many more users who could potentially answer (but also, as choward points out, more questions that could drown out yours).

  • $\begingroup$ I'm shocked to learn that (mathematical) optimization is best here. Great to know though! $\endgroup$ Jul 8 '16 at 18:02
  • $\begingroup$ What about deep learning and optimization methods related to it? Consider for the sake of an example consider the momentum algorithm. Should I ask about its properties here to understand its role in solving non-convex problems? $\endgroup$ Jul 8 '16 at 18:04
  • $\begingroup$ I was also particularly interested in properties of batch normalization (deep learning optimization algorithm). Is that to be asked here or on Cross Validate? $\endgroup$ Jul 8 '16 at 18:04

You must log in to answer this question.

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