Syntactically Guided Neural Machine Translation
This work addresses translation quality and scalability issues for NMT practitioners, but it is incremental as it builds on existing methods.
The authors tackled the problem of improving neural machine translation (NMT) by integrating hierarchical phrase-based statistical machine translation (SMT) lattices, resulting in gains over both Hiero and NMT decoding alone with practical advantages for large vocabularies.
We investigate the use of hierarchical phrase-based SMT lattices in end-to-end neural machine translation (NMT). Weight pushing transforms the Hiero scores for complete translation hypotheses, with the full translation grammar score and full n-gram language model score, into posteriors compatible with NMT predictive probabilities. With a slightly modified NMT beam-search decoder we find gains over both Hiero and NMT decoding alone, with practical advantages in extending NMT to very large input and output vocabularies.