For a dataset with multi-label judgement, e.g., coco dataset but where we only want to predict the most-possible label. There're multiple ways, for example : 1) train as a multi-label learning(each label as a binary-task) and predict as a multi-class problem using sigmoid and predict the one with largest sigmoid score; 2) train as a multi-class learning (e.g., change the data) and predict in a multi-class way.

My question is for such problem, how to choose which methods to use ? And what is the standard handling for such problems ?

It seems not much literature to support the related connection. It would be much appreciated to see if there's reference to point to related research or some study on more general connection between multi-label learning and multi-class learning in both theory and practice.

  • Multi-Label Learning In Multi-label learning, one training example is associated with multiple class labels simultaneously. The multi-label learner induces a function that is able to assign multiple proper labels (from a given label set) to unseen instances

    Best example of multi-label is genres of a movie like Comedy, Drama, Romance, Action, Adventure, Animation, Family, Thriller

  • Multi-Class Learning is the problem of classifying instances into one of three or more classes.

    Best example of multi-class is Motion Picture Association film rating system like G(General Audiences), PG(Parental Guidance Suggested), PG-13(Parents Strongly Cautioned), R(Restricted)

In Simpler terms, multi-class classification is when each input will have only one output class, but in multi-label classification each input can have multi-output classes.


  • $\begingroup$ The question is about the connection, raised from a conversion problem, instead of definition of each other, or the difference $\endgroup$
    – exteral
    Oct 14 at 23:32
  • $\begingroup$ @exteral I can see you have recently updated your question, which is not something I predicted. If I come across the mentioned problem will update with relevant answer. $\endgroup$ Oct 15 at 2:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.