Parallel Blockwise Knowledge Distillation for Deep Neural Network Compression | IEEE Journals & Magazine | IEEE Xplore