4 ms·Regularizing Neural Networks by Penalizing Confident Output Distributions2 points by tonybeltramelli 10y agotonybeltramelli 10y agoImprove exploration in reinforcement learning, acts as a regularizer in supervised learning.