Improving the interpretability of deep neural networks with stimulated learning | IEEE Conference Publication | IEEE Xplore