About the formula 3 in the paper.

nayeemrizve / ups

"In Defense of Pseudo-Labeling: An Uncertainty-Aware Pseudo-label Selection Framework for Semi-Supervised Learning" by Mamshad Nayeem Rizve, Kevin Duarte, Yogesh S Rawat, Mubarak Shah (ICLR 2021)

MIT License

231 stars 40 forks source link

About the formula 3 in the paper. #1

Closed Kouuh closed 3 years ago

Kouuh commented 3 years ago

Whether the first y in the formula is a hard pseudo-label generated by the network prediction.

nayeemrizve commented 3 years ago

Yes

ghost commented 3 years ago

So in here, why you only use 1 instead of (1-sim y).

Do I miss something?