In response to a common criticism of deep learning which is that you’ve got this big black box network, you trained it on gradient ascent, you get a good number but we don’t trust the network because we don’t understand as people why it’s making the decision that it’s making.