fil-mp opened this issue 4 years ago · 2 comments
Why do the authors include the labeled examples in the unlabeled dataset?
I think it is because of the stability of learning unlabeled data.
There is also the meaning of augmentation of labeled data.