Fetching the paper…

Regularizing Neural Networks by Penalizing Confident Output Distributions · Around