I often need to discern what the State of the Art (SOTA) is for certain Machine Learning Applications, which I currently ask on the Fast.ai forums. For example:

I don't think these are on-topic, because the question essentially boils down to "what's out there?" which isn't very answerable and seems more suitable for a forum and/or wiki. However, I wanted to check my conclusion with the community before self-filtering.

  • 7
    $\begingroup$ I'm afraid that with "what's deep learning SOTA for X" we would have to update the answers every week... $\endgroup$
    – Tim
    Commented Mar 15, 2019 at 15:40
  • 3
    $\begingroup$ Somewhat related is: stats.meta.stackexchange.com/questions/5540/… I believe that is very important for such questions to describe a clear scope, and this is a bit of a grey area. What is expected to be a sufficient answer? A question that is phrased as 'how does X deal with Y?' is already more specific than 'what are (latest) examples of X dealing with Y' (how big of a list is expected, does it need to be exhaustive, just provide a few of the most recent/important, how novel does it need to be to classify as sota, etc?) $\endgroup$ Commented Mar 15, 2019 at 16:38
  • $\begingroup$ No one objected to my question about SOTA for some image benchmarks stats.stackexchange.com/questions/363640/… $\endgroup$
    – Sycorax Mod
    Commented Mar 23, 2019 at 16:36

1 Answer 1


Both of those examples seem to me to be well researched, & narrow enough in scope for our Q&A site. "What's state-of-the-art in supervised learning", on the other hand, would, I think, be too broad.

  • 5
    $\begingroup$ In other words, specific problems presented with sufficient context can be considered on-topic? I'm just trying to rephrase this to make sure I understand. $\endgroup$
    – Seanny123
    Commented Mar 15, 2019 at 16:17
  • $\begingroup$ @Seanny123: Yes. The 2nd example could perhaps do with some clarification: what kind of explanation are you looking for? - global or local? $\endgroup$ Commented Mar 15, 2019 at 18:00

You must log in to answer this question.

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