Interpretations are Useful Penalizing Explanations to Align Neural Networks with Prior Knowledge

InterpretationsareUseful:PenalizingExplanationstoAlignNeuralNetworkswithPriorKnowledgeLauraRieger1ChandanSingh2W.JamesMurdoch3BinYu23Abstractimproveamodel.ForanexplanationofadeeplearningmodeltoPred...