Chunk-Based Bi-Scale Decoder for Neural Machine Translation
This work addresses translation quality for machine translation systems, offering a novel decoder design that is incremental in improving existing methods.
The paper tackles the problem of neural machine translation by proposing a chunk-based bi-scale decoder that hierarchically translates sentences from chunks to words, resulting in significant performance improvements over the state-of-the-art model.
In typical neural machine translation~(NMT), the decoder generates a sentence word by word, packing all linguistic granularities in the same time-scale of RNN. In this paper, we propose a new type of decoder for NMT, which splits the decode state into two parts and updates them in two different time-scales. Specifically, we first predict a chunk time-scale state for phrasal modeling, on top of which multiple word time-scale states are generated. In this way, the target sentence is translated hierarchically from chunks to words, with information in different granularities being leveraged. Experiments show that our proposed model significantly improves the translation performance over the state-of-the-art NMT model.