The tag has no usage information. I do not know what it is for, anybody knows, and can provide a wiki?

  • 8
    $\begingroup$ Stochastic Gradient Descent? $\endgroup$ Commented Nov 5, 2017 at 18:30
  • 1
    $\begingroup$ The tag seems to have been created by @avocado here. This is an example of our perennial problem w/ people using acronyms w/o defining them; "SVD" is never defined in either the question or the answer. $\endgroup$ Commented Nov 5, 2017 at 19:14
  • 3
    $\begingroup$ @MatthewDrury, that seems to be the answer, why not make it an 'official' answer here, & create an excerpt for the tag? $\endgroup$ Commented Nov 5, 2017 at 19:17

1 Answer 1


I wrote the following wiki excerpt:

Stochastic gradient descent (SGD) is a variant of gradient descent where only a small subset ("mini-batch") of training examples is used to compute the gradient on each iteration.

Feel free to improve.

It might make sense to create a synonym [stochastic-gradient-descent]$\to$.

Update: done.

  • 1
    $\begingroup$ Please create the synonym! $\endgroup$ Commented Nov 6, 2017 at 11:16
  • 2
    $\begingroup$ @kjetilbhalvorsen, amoeba, I created the synonym here. $\endgroup$ Commented Nov 6, 2017 at 12:46
  • 4
    $\begingroup$ @gung maybe we should swap the synonyms so the full form is the basic tag and sgd is a synonym? PS I created meta thread for this: stats.meta.stackexchange.com/questions/5022/… $\endgroup$
    – Tim
    Commented Nov 6, 2017 at 15:22
  • $\begingroup$ That's potentially fine w/ me, @Tim. My 1st thought is that people who ask / answer about this seem to have preferred the short version, so we might take that into account. It might also be worth discussing. What do amoeba, Kjetil, & Matt think, eg? $\endgroup$ Commented Nov 6, 2017 at 15:28
  • $\begingroup$ @kjetilbhalvorsen, see comments above. Do you have a preference? $\endgroup$ Commented Nov 6, 2017 at 15:28
  • $\begingroup$ @gung I have no preference. $\endgroup$ Commented Nov 6, 2017 at 15:30
  • 2
    $\begingroup$ I'd prefer the full version be primary, and I'd prefer that was our general rule! $\endgroup$ Commented Nov 6, 2017 at 15:35

You must log in to answer this question.

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