4 ms·Improving Interpretability of Deep Neural Networks with Semantic Information2 points by Katydid 10y ago