Real-world uses

Created for this library

1. 1.

   A click-prediction team uses cross-entropy loss to train its binary classifier because it aligns with the calibrated probability the bidder needs.

2. 2.

   A medical screening team uses cross-entropy with class weights so the rare positive class drives more of the loss during training.

3. 3.

   A search ranking team uses cross-entropy on a click target as one of several losses in its multi-objective ranker.