Exploiting Semantics in Neural Machine Translation with Graph\n Convolutional Networks

2018/04/23 by Diego Marcheggiani, Marcheggiani, Diego, Jasmijn Bastings +3 · 1 citation
Computer Science · #Natural Language Processing Techniques #Topic Modeling #Text Readability and Simplification

paper · pdf · doi:10.48550/arxiv.1804.08313

Abstract

Semantic representations have long been argued as potentially useful for\nenforcing meaning preservation and improving generalization performance of\nmachine translation methods. In this work, we are the first to incorporate\ninformation about predicate-argument structure of source sentences (namely,\nsemantic-role representations) into neural machine translation. We use Graph\nConvolutional Networks (GCNs) to inject a semantic bias into sentence encoders\nand achieve improvements in BLEU scores over the linguistic-agnostic and\nsyntax-aware versions on the English--German language pair.\n

Cited by

• LLMs Can Also Do Well! Breaking Barriers in Semantic Role Labeling via Large Language Models

Related

• Graph Convolutional Encoders for Syntax-aware Neural Machine Translation
• Neural Machine Translation: A Review and Survey
• Document Graph for Neural Machine Translation
• Neural Machine Translation with Source-Side Latent Graph Parsing
• Generative Neural Machine Translation