Attention With Sparsity Regularization for Neural Machine Translation and Summarization | IEEE Journals & Magazine | IEEE Xplore