Regularizing Neural Networks by Penalizing Confident Output Distributions

Improve exploration in reinforcement learning, acts as a regularizer in supervised learning.