I'm new to deep learning. I was wondering what's the relationship between a deep model complexity (e.g. total number of parameters, or depth) and the dataset size?

Assuming I want to do a binary classification with 10K data for a problem like fire detection. How should I know what complexity I should go for?

  • 1
    $\begingroup$ In general, this is may not be an easy question to answer. The branch of artificial intelligence that studies these issues is called computational (and statistical) learning theory. See also cs.stackexchange.com/q/75327/20691. I will eventually try to answer this question, but I don't know when. $\endgroup$
    – nbro
    Dec 11, 2019 at 23:02
  • $\begingroup$ Thanks. Yeah try to write something so i can give u some credit. $\endgroup$
    – Mary
    Dec 12, 2019 at 3:23
  • $\begingroup$ A recent question (ai.stackexchange.com/questions/39176/…) seems related to this one. $\endgroup$ Feb 21, 2023 at 17:17


You must log in to answer this question.