Comparable corpora and computer-assisted translation

305 Pages·2014·2.182 MB·English

by  Delpech, Estelle Maryline

W689-Delpech.qxp_Layout 1 12/06/2014 08:30 Page 1
COGNITIVE SCIENCE AND KNOWLEDGE MANAGEMENT SERIES
Computer-assisted translation (CAT) has always used translation E
s
memories, which require the translator to have a corpus of previous te
translations that the CAT software can use to generate bilingual lle

M
lexicons. This can be problematic when the translator does not have
a
such a corpus, for instance, when the text belongs to an emerging field. ry
To solve this issue, CAT research has looked into the leveraging of lin Comparable Corpora
e
comparable corpora, i.e. a set of texts, in two or more languages, which D
e
deal with the same topic but are not translations of one another. lp and Computer-assisted
e
c
This work had two primary objectives. The first is to assess the input of h

lexicons extracted from comparable corpora in the context of a
specialized human translation task. The second objective is to identify Translation
bilingual-lexicon-extraction methods which best match the translators’ C
o
needs, determining the current limits of these techniques and
m
suggesting improvements. The author focuses, in particular, on the
p
identification of fertile translations, the management of multiple a
r
morphological structures, and the ranking of candidate translations. a
b
The experiments are carried out on two language pairs (English–French le
and English–German) and on specialized texts dealing with breast C
cancer. This research puts significant emphasis on applicability – o Estelle Maryline Delpech
r
methodological choices are guided by the needs of the final users. This p
o
book is organized in two parts: the first part presents the applicative and r
a
scientific context of the research, and the second part is given over to

a
efforts to improve compositional translation. n
d
The research work presented in this book received the PhD Thesis
C
award 2014 from the French association for natural language o
m
processing (ATALA).
p
u
t
e
r
-
a
s
s
i
s
Estelle Maryline Delpech holds a PhD in Computer Science from the t
e
University of Nantes in France, where she specialized in natural d

language processing and computer-aided translation. She is currently T
r
Chief Scientist at Nomao, a web and mobile app search engine a
n
company. Her research interests include multilingualism, computational s
l
linguistics, information extraction and data integration. a
t
i
o
n
Z(7ib8e8-CBGIJB(
www.iste.co.uk

Comparable Corpora and Computer-assisted Translation
To Elia
Series Editor
Narendra Jussien
Comparable Corpora and
Computer-assisted Translation
Estelle Maryline Delpech
Firstpublished2014inGreatBritainandtheUnitedStatesbyISTELtdandJohnWiley&Sons,Inc.
Apart from any fair dealing for the purposes of research or private study, or criticism or review, as
permittedundertheCopyright,DesignsandPatentsAct1988,thispublicationmayonlybereproduced,
storedortransmitted,inanyformorbyanymeans,withthepriorpermissioninwritingofthepublishers,
or in the case of reprographic reproduction in accordance with the terms and licenses issued by the
CLA. Enquiries concerning reproduction outside these terms should be sent to the publishers at the
undermentionedaddress:
ISTELtd JohnWiley&Sons,Inc.
27-37StGeorge’sRoad 111RiverStreet
LondonSW194EU Hoboken,NJ07030
UK USA
www.iste.co.uk www.wiley.com
©ISTELtd2014
TherightsofEstelleMarylineDelpechtobeidentifiedastheauthorofthisworkhavebeenassertedby
herinaccordancewiththeCopyright,DesignsandPatentsAct1988.
LibraryofCongressControlNumber:2014936484
BritishLibraryCataloguing-in-PublicationData
ACIPrecordforthisbookisavailablefromtheBritishLibrary
ISBN978-1-84821-689-1
PrintedandboundinGreatBritainbyCPIGroup(UK)Ltd.,Croydon,SurreyCR04YY
Table of Contents
ACKNOWLEDGMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix
INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi
PART1.APPLICATIVEAND SCIENTIFICCONTEXT . . . . . . . . . . . . . 1
CHAPTER 1. LEVERAGING COMPARABLE CORPORA FOR COMPUTER-
ASSISTEDTRANSLATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2. From the beginnings of machine translation to comparable corpora
processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2.1. Thedawnofmachinetranslation . . . . . . . . . . . . . . . . . . . 3
1.2.2. Thedevelopmentofcomputer-assistedtranslation . . . . . . . . . . 5
1.2.3. Drawbacksofparallelcorporaandadvantagesofcomparablecorpora 7
1.2.4. Difficultiesoftechnicaltranslation . . . . . . . . . . . . . . . . . . 9
1.2.5. Industrialcontext . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.3. Termalignmentfromcomparablecorpora: astate-of-the-art . . . . . . 15
1.3.1. Distributionalapproachprinciple . . . . . . . . . . . . . . . . . . . 15
1.3.2. Termalignmentevaluation . . . . . . . . . . . . . . . . . . . . . . . 18
1.3.3. Improvementandvariantsofthedistributionalapproach . . . . . . 20
1.3.4. Theinfluencedataandparametersonalignmentquality . . . . . . 28
1.3.5. Limitsofthedistributionalapproach . . . . . . . . . . . . . . . . . 30
1.4. CATsoftwareprototypeforcomparablecorporaprocessing . . . . . . 32
1.4.1. Implementationofatermalignmentmethod . . . . . . . . . . . . . 32
1.4.2. Terminologicalrecordsextraction . . . . . . . . . . . . . . . . . . . 36
vi ComparableCorporaandComputer-assistedTranslation
1.4.3. Lexiconconsultationinterface . . . . . . . . . . . . . . . . . . . . . 38
1.5. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
CHAPTER2.USER-CENTERED EVALUATIONOF LEXICONS
EXTRACTEDFROM COMPARABLE CORPORA . . . . . . . . . . . . . . . . . 41
2.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
2.2. Translationqualityevaluationmethodologies. . . . . . . . . . . . . . . 42
2.2.1. Machinetranslationevaluation. . . . . . . . . . . . . . . . . . . . . 42
2.2.2. Humantranslationevaluation . . . . . . . . . . . . . . . . . . . . . 46
2.2.3. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
2.3. Designandexperimentationofauser-centeredevaluation . . . . . . . 50
2.3.1. Methodologicalaspects . . . . . . . . . . . . . . . . . . . . . . . . . 50
2.3.2. Experimentationprotocol. . . . . . . . . . . . . . . . . . . . . . . . 54
2.3.3. Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
2.4. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
CHAPTER3.AUTOMATICGENERATIONOF TERM TRANSLATIONS . . . 67
3.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67
3.2. Compositionalapproaches . . . . . . . . . . . . . . . . . . . . . . . . . 67
3.2.1. Compositionaltranslationprinciple . . . . . . . . . . . . . . . . . . 68
3.2.2. Polylexicalunitscompositionaltranslation . . . . . . . . . . . . . . 70
3.2.3. Monolexicalunitscompositionaltranslation . . . . . . . . . . . . . 75
3.2.4. Candidatetranslationfiltering . . . . . . . . . . . . . . . . . . . . . 81
3.3. Data-drivenapproaches . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
3.3.1. Analogy-basedtranslation . . . . . . . . . . . . . . . . . . . . . . . 85
3.3.2. Rewritingruleslearning . . . . . . . . . . . . . . . . . . . . . . . . 87
3.3.3. Dealingwithmorphologicalvariation . . . . . . . . . . . . . . . . . 88
3.4. Evaluationoftermtranslatorgenerationmethods . . . . . . . . . . . . 91
3.5. Researchperspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
PART2.CONTRIBUTIONSTOCOMPOSITIONALTRANSLATION . . . . . 99
CHAPTER4.MORPH-COMPOSITIONALTRANSLATION:
METHODOLOGICALFRAMEWORK . . . . . . . . . . . . . . . . . . . . . . . 101
4.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101
4.2. Morpho-compositionaltranslationmethod . . . . . . . . . . . . . . . . 101
4.2.1. Scientificpositioning . . . . . . . . . . . . . . . . . . . . . . . . . . 102
4.2.2. Definitionsandterminology . . . . . . . . . . . . . . . . . . . . . . 105
4.2.3. Underlyingassumptions . . . . . . . . . . . . . . . . . . . . . . . . 108
4.2.4. Advantagesoftheproposedapproachforprocessingcomparable
corpora . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
4.3. Addressedissuesandcontributions . . . . . . . . . . . . . . . . . . . . 110
4.3.1. Generatingfertiletranslations . . . . . . . . . . . . . . . . . . . . . 110
Contents vii
4.3.2. Dealingwithdiversemorphologicalstructures . . . . . . . . . . . . 113
4.3.3. Candidatetranslationsranking . . . . . . . . . . . . . . . . . . . . . 116
4.4. Evaluationmethodology . . . . . . . . . . . . . . . . . . . . . . . . . . 117
4.4.1. Apriorireference . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
4.4.2. Aposteriorireference . . . . . . . . . . . . . . . . . . . . . . . . . . 120
4.5. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 121
CHAPTER5.EXPERIMENTALDATA . . . . . . . . . . . . . . . . . . . . . . . 123
5.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
5.2. Comparablecorpora . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 123
5.3. Sourceterms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125
5.4. Referencedatafortranslationgenerationevaluation . . . . . . . . . . . 126
5.4.1. Apriorireference . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
5.4.2. Aposteriorireference . . . . . . . . . . . . . . . . . . . . . . . . . . 129
5.5. Translationrankingtrainingandevaluationdata . . . . . . . . . . . . . 131
5.6. Linguisticresources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131
5.6.1. Generallanguagebilingualdictionary. . . . . . . . . . . . . . . . . 131
5.6.2. Thesaurus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
5.6.3. Boundmorphemestranslationtable . . . . . . . . . . . . . . . . . . 132
5.6.4. Lexiconforworddecomposition . . . . . . . . . . . . . . . . . . . 133
5.6.5. Morphologicalfamilies . . . . . . . . . . . . . . . . . . . . . . . . . 134
5.6.6. Dictionaryofcognates . . . . . . . . . . . . . . . . . . . . . . . . . 135
5.7. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
CHAPTER 6. FORMALIZATION AND EVALUATION OF CANDIDATE
TRANSLATION GENERATION . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
6.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139
6.2. Translationgenerationalgorithm . . . . . . . . . . . . . . . . . . . . . . 139
6.2.1. Decomposition. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 142
6.2.2. Translation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 144
6.2.3. Recomposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 145
6.2.4. Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 146
6.3. Morphologicalsplittingevaluation. . . . . . . . . . . . . . . . . . . . . 147
6.4. Translationgenerationevaluation . . . . . . . . . . . . . . . . . . . . . 148
6.4.1. Referencedataandevaluationmeasures . . . . . . . . . . . . . . . 148
6.4.2. Modelgenericityinfluence . . . . . . . . . . . . . . . . . . . . . . . 152
6.4.3. Linguisticresourcesinfluence . . . . . . . . . . . . . . . . . . . . . 156
6.4.4. Fallbackstrategyinfluence . . . . . . . . . . . . . . . . . . . . . . . 159
6.4.5. Fertiletranslationsinfluence . . . . . . . . . . . . . . . . . . . . . . 160
6.4.6. Popularsciencecorpusinfluence . . . . . . . . . . . . . . . . . . . 165
6.4.7. Qualitativeanalysis . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
6.5. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
6.5.1. Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 173
6.5.2. Researchperspectives . . . . . . . . . . . . . . . . . . . . . . . . . . 176
viii ComparableCorporaandComputer-assistedTranslation
CHAPTER 7. FORMALIZATION AND EVALUATION OF CANDIDATE
TRANSLATION RANKING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
7.1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
7.2. Rankingcriteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
7.2.1. Contextsimilarity . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179
7.2.2. Candidatetranslationfrequency . . . . . . . . . . . . . . . . . . . . 180
7.2.3. Parts-of-speechtranslationprobability . . . . . . . . . . . . . . . . 180
7.2.4. Componentstranslationmode . . . . . . . . . . . . . . . . . . . . . 181
7.3. Criteriacombination. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
7.3.1. Valuestandardization . . . . . . . . . . . . . . . . . . . . . . . . . . 184
7.3.2. Linearcombination . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
7.3.3. Learning-to-rankmodel. . . . . . . . . . . . . . . . . . . . . . . . . 186
7.4. Evaluation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
7.4.1. Referencedataandevaluationmeasures . . . . . . . . . . . . . . . 187
7.4.2. Basesofcomparison . . . . . . . . . . . . . . . . . . . . . . . . . . 188
7.4.3. Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 189
7.5. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 191
7.5.1. Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 192
7.5.2. Researchperspectives . . . . . . . . . . . . . . . . . . . . . . . . . . 196
CONCLUSIONANDPERSPECTIVES . . . . . . . . . . . . . . . . . . . . . . . 199
PART3.APPENDICES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
APPENDIX1.MEASURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 207
APPENDIX2.DATA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 215
APPENDIX3.COMPARABLE CORPORALEXICONS
CONSULTATION INTERFACE . . . . . . . . . . . . . . . . . . . . . . . . . . . 261
LISTOFTABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 265
LISTOFFIGURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271
LISTOFALGORITHMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 273
LISTOFEXTRACTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 275
BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
INDEX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289