Exploiting Semantics in Neural Machine Translation with Graph Convolutional Networks

Open Access
Authors
Publication date 2018
Host editors
  • M. Walker
  • H. Ji
  • A. Stent
Book title NAACL-HLT 2018 : The 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies
Book subtitle proceedings of the conference : June 1-June 6, 2018, New Orleans, Louisiana
ISBN (electronic)
  • 9781948087292
Event 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies
Volume | Issue number 2
Pages (from-to) 486–492
Number of pages 7
Publisher Stroudsburg, PA: The Association for Computational Linguistics
Organisations
  • Interfacultary Research - Institute for Logic, Language and Computation (ILLC)
Abstract Semantic representations have long been argued as potentially useful for enforcing meaning preservation and improving generalization performance of machine translation methods. In this work, we are the first to incorporate information about predicate-argument structure of source sentences (namely, semantic-role representations) into neural machine translation. We use Graph Convolutional Networks (GCNs) to inject a semantic bias into sentence encoders and achieve improvements in BLEU scores over the linguistic-agnostic and syntax-aware versions on the English–German language pair.
Document type Conference contribution
Note Later version also available. - With supplementary notes.
Language English
Published at https://doi.org/10.18653/v1/N18-2078
Downloads
N18-2078v2 (Other version)
Supplementary materials
Permalink to this page
Back