Towards Neural Phrase-based Machine Translation

Po-Sen Huang, Chong Wang, Sitao Huang, Dengyong Zhou, Li Deng
6/17/2017
cs.CLstat.ML

Official code: posenhuang/NPMT (177 stars, last commit 2018-07-11).

Abstract

In this paper, we present Neural Phrase-based Machine Translation (NPMT). Our method explicitly models the phrase structures in output sequences using Sleep-WAke Networks (SWAN), a recently proposed segmentation-based sequence modeling method. To mitigate the monotonic alignment requirement of SWAN, we introduce a new layer to perform (soft) local reordering of input sequences. Different from existing neural machine translation (NMT) approaches, NPMT does not use attention-based decoding mechanisms. Instead, it directly outputs phrases in a sequential order and can decode in linear time. Our experiments show that NPMT achieves superior performances on IWSLT 2014 German-English/English-German and IWSLT 2015 English-Vietnamese machine translation tasks compared with strong NMT baselines. We also observe that our method produces meaningful phrases in output languages.

View on arXivView PDF

Code Implementations(1)

posenhuang/NPMTOfficial100%

Towards Neural Phrase-based Machine Translation

17727C++, LuaJan 29, 20188 years agoNOASSERTION
fairseqluamachine-translationneural-machine-translationnpmt+3 more

Similar Papers

Neural Semantic Parsing by Character-based Translation: Experiments with Abstract Meaning Representations

83% match1 repo472017

Levenshtein Transformer

82% match1 repo32.2k2019

Discrete Autoencoders for Sequence Models

82% match1 repo17.5k2018

Empower Sequence Labeling with Task-Aware Neural Language Model

82% match1 repo8442017

Neural Machine Translation with Dynamic Graph Convolutional Decoder

82% match2023

Training Tips for the Transformer Model

82% match1 repo17.5k2018

Cite this paper

@article{huang2017towards,
  title  = {Towards Neural Phrase-based Machine Translation},
  author = {Po-Sen Huang and Chong Wang and Sitao Huang and Dengyong Zhou and Li Deng},
  year   = {2017},
  eprint = {1706.05565},
  archivePrefix = {arXiv},
  url    = {https://arxiv.org/abs/1706.05565}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/1706.05565.svg)](https://codeofpaper.com/paper/1706.05565)

Discussion