Chapter 2. ReGap
-
Carlos Manuel Hidalgo-Ternero
Abstract
This research presents ReGap, a text-preprocessing algorithm for the automatic token-based identification and conversion of discontinuous multiword expressions (MWEs) into their canonical state, i.e., their continuous form, as a means to optimise neural machine translation (NMT) systems. To this end, an experiment with flexible verb-noun idiomatic constructions (VNICs) is conducted in order to assess to what extent ReGap can enhance the performance of the most robust NMT system to date, DeepL, under the challenge of MWE discontinuity in the Spanish-into-English and the Spanish-into-German directionalities. In this regard, the promising results yielded for VNICs will shed some light on new avenues for enhancing MWE‑aware NMT systems.
Abstract
This research presents ReGap, a text-preprocessing algorithm for the automatic token-based identification and conversion of discontinuous multiword expressions (MWEs) into their canonical state, i.e., their continuous form, as a means to optimise neural machine translation (NMT) systems. To this end, an experiment with flexible verb-noun idiomatic constructions (VNICs) is conducted in order to assess to what extent ReGap can enhance the performance of the most robust NMT system to date, DeepL, under the challenge of MWE discontinuity in the Spanish-into-English and the Spanish-into-German directionalities. In this regard, the promising results yielded for VNICs will shed some light on new avenues for enhancing MWE‑aware NMT systems.
Chapters in this book
- Prelim pages i
- Table of contents v
- Preface vii
-
Section 1. Computational treatment of multiword units
- Chapter 1. Multi-word units in neural machine translation 2
- Chapter 2. ReGap 18
- Chapter 3. Evaluating the Italian-English machine translation quality of MWUs in the domain of archaeology 40
- Chapter 4. Post-editing neural machine translation in specialised languages 57
- Chapter 5. Evaluating a bracketing protocol for multiword terms 79
-
Section 2. Corpus-based and linguistic studies in phraseology
- Chapter 6. Suggestions for a new model of functional phraseme categorization for applied purposes 104
- Chapter 7. Verb collocations and their semantics in the specialized language of science 124
- Chapter 8. Negative–positive adjective pairing in travel journalism in English, Italian, and Polish 141
- Chapter 9. The middle construction and some machine translation issues 156
- Chapter 10. Semantic annotation of named rivers and its application for the prediction of multiword-term bracketing 173
- Chapter 11. Irony in American-English tweets 197
- Chapter 12. A comprehensive Japanese MWE lexicon 218
- Chapter 13. Ontology-based formalisation of Italian clitic verbal MWEs 243
- Index 263
Chapters in this book
- Prelim pages i
- Table of contents v
- Preface vii
-
Section 1. Computational treatment of multiword units
- Chapter 1. Multi-word units in neural machine translation 2
- Chapter 2. ReGap 18
- Chapter 3. Evaluating the Italian-English machine translation quality of MWUs in the domain of archaeology 40
- Chapter 4. Post-editing neural machine translation in specialised languages 57
- Chapter 5. Evaluating a bracketing protocol for multiword terms 79
-
Section 2. Corpus-based and linguistic studies in phraseology
- Chapter 6. Suggestions for a new model of functional phraseme categorization for applied purposes 104
- Chapter 7. Verb collocations and their semantics in the specialized language of science 124
- Chapter 8. Negative–positive adjective pairing in travel journalism in English, Italian, and Polish 141
- Chapter 9. The middle construction and some machine translation issues 156
- Chapter 10. Semantic annotation of named rivers and its application for the prediction of multiword-term bracketing 173
- Chapter 11. Irony in American-English tweets 197
- Chapter 12. A comprehensive Japanese MWE lexicon 218
- Chapter 13. Ontology-based formalisation of Italian clitic verbal MWEs 243
- Index 263
