Towards Global Explanations of Convolutional Neural Networks With Concept Attribution | IEEE Conference Publication | IEEE Xplore