PHAED: A Speaker-Aware Parallel Hierarchical Attentive Encoder-Decoder Model for Multi-Turn Dialogue Generation | IEEE Journals & Magazine | IEEE Xplore