Regularizing Neural Networks by Penalizing Confident Output Distributions

Pereyra, Gabriel, Tucker, George, Chorowski, Jan · arXiv · 2017