Neural Hidden Markov Model for Machine Translation
Weiyue Wang, Derui Zhu, Tamer Alkhouli, Zixuan Gan, Hermann Ney · 2018
This work aims to investigate alternative neural machine translation (NMT) approaches and thus proposes a neural hidden Markov model (HMM) consisting of neural network-based alignment and lexicon models.The neural models make use of encoder and decoder components, but drop the attention component.The training is end-to-end and the standalone decoder is able to provide comparable performance with the state-of-the-art attention-based models on three different translation tasks.