MétaCan
Menu
Retour à la cohorte
Enregistrement W2268831157

Corpus-based Machine Translation: Its Current Development and Perspectives

2015· article· en· W2268831157 sur OpenAlexaboutno aff
Zhou Da-jun, Yun Wang

Notice bibliographique

RevueInternational Forum of Teaching and Studies · 2015
Typearticle
Langueen
DomaineComputer Science
ThématiqueNatural Language Processing Techniques
Établissements canadiensnon disponible
Organismes subventionnairesnon disponible
Mots-clésMachine translationComputer scienceArtificial intelligenceNatural language processingExample-based machine translationCorpus linguisticsParallel corporaComputer-assisted translationMachine translation software usabilityRule-based machine translationText corpusComputational linguistics
DOInon disponible

Résumé

récupéré en direct d'OpenAlex

A Review of Corpus-based Machine TranslationCorpus and Machine TranslationCorpus is a large-scale database with tremendous collective linguistic information in real use, which is provided for retrieval by computers for research. The first corpus was established in Brown University, U.S.A., in the late 1960s (Zhang & Zhang, 2010, p. 55). Much progress has been made in corpus research and application in the past decades. Current studies of parallel or multi-language corpus can be categorized into three aspects: the first is the alignment technology of parallel language material, with various strategies and approaches provided by scholars and with numbers of programs and tools of alignment parallel or multi-language material; the second is the application of parallel language materials, such as statistic-based machine translation, example-based machine translation, and parallel language dictionary compiling; the third refers to issues of parallel language corpus design and management, and its material collection and coding (Chang, et al., 2003, p. 28).Corpus used in translation is one of the focuses of corpus application research. Machine translation (MT) is a technology to translate from one natural language in character or speech form into another by means of computer programs (Zhao & Liu, 2010, p. 36). MT was initiated in the 1950s and entered into a prosperous period in the end of 1980s, which characterizes practicality of many translation systems in various fields. An English-French translation system TAUMMETEO developed by the University of Montreal, Canada, in 1976, is a typical example, which can provide high-quality translation of weather forecast (Shao, 2010, p. 28). A typical MT system adopts a transfer-based translation strategy, which consists of 3 procedures: 1) analyzing a source language and form representation of the language; 2) transforming representation of the source language into that of the target language; 3) generating the source language translation version from the target language representation (He, 2007, p. 191). Traditional MT has two defects: one is that traditional MT regards words as its basic translation unit; that is, the machine first segments sentences of a source language into words, which are transformed into those of target language, and then those words are connected according to grammatical structure rules of the target language; the other is that traditional MT does not pay much attention to contexts. Peer-to-peer studies of corpus-based MT do attempt to get over the defects of the traditional MT systems and improve efficiencies and accuracy of the MT systems.With the development over more than 50 years, MT systems have performed certain functions in some fields. However, the current systems have not reached the effect of translation as expected. In its earlier period, MT research was conducted from the viewpoint of natural linguistics, thus creating MT systems based on such linguistic rules as lexical rule, syntactic analysis rule, transformation rules, and target language generation rules. As these rules were summarized and developed from experiences of linguists, there exist some deficiencies in the analytic rules. For instance, manual writing of those rules demands a large quantity of workload, and the rules are too subjective to keep consistency (Wang, 2003, p. 33). Since 1989, MT has entered a new stage in which corpus methods are introduced to the rule-based technology, including statistic-based and example-based methods, and the method of turning corpus into linguistic knowledge bank through language material processing, etc. (Feng, 2010, p. 28). The past years have seen the prominent achievement in MT systems.Corpus Used in Machine TranslationThere are three kinds of corpus concerning MT: parallel corpus, multi-language corpus, and comparable corpus. Parallel corpus collects original text of a language and its translating text of another language. …

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.

Comment cette classification a été obtenuedéplier

Prédiction distillée sur la base complète

Imitation des enseignants

Ni prévalence calibrée, ni vérité terrain. Validation humaine à venir. Apprise à partir de 10 348 étiquettes directes de Codex et de 10 348 étiquettes directes de Gemma. Le mode candidate est l'union des têtes enseignantes seuillées; le consensus est leur intersection. Ces sorties portent le statut machine_predicted_unvalidated et ne sont ni des étiquettes humaines ni des étiquettes directes de modèles de pointe.

score de la tête « metaresearch » (Codex)0,000
score de la tête « metaresearch » (Gemma)0,000
Version: codex-gemma-dda1882f352aStatut de validation: machine_predicted_unvalidated
Catégories candidatesaucune
Catégories consensuellesaucune
DomaineSignal candidat: aucune · Signal consensuel: aucune
Devis d'étudeSignal candidat: Autre devis · Signal consensuel: aucune
GenreSignal candidat: Empirique · Signal consensuel: aucune
Score de désaccord entre enseignants0,856
Score d'incertitude au seuil0,256

Scores Codex et Gemma par catégorie

CatégorieCodexGemma
Métarecherche0,0000,000
Méta-épidémiologie (sens strict)0,0000,000
Méta-épidémiologie (sens large)0,0000,000
Bibliométrie0,0000,000
Études des sciences et des technologies0,0000,000
Communication savante0,0000,000
Science ouverte0,0000,000
Intégrité de la recherche0,0000,000
Charge utile insuffisante (le modèle a refusé de juger)0,0000,000

Scores machine (provisoires)

Les deux têtes enseignantes du modèle étudiant, lues sur ce travail. Un score ordonne la base pour la relecture; il n'affirme jamais une catégorie, et le statut de validation accompagne chaque rangée tel quel.

Scores de référence d'un modèle non mature (critères de maturité non atteints, 7 itérations). Un score ordonne; il n'affirme jamais une catégorie.

Tête enseignante Opus0,090
Tête enseignante GPT0,362
Écart entre enseignants0,272 · la distance entre les deux têtes enseignantes sur ce seul travail
Statut de validationscore_only:v0-immature-baseline · tel quel depuis la passe de notation : score_only signifie que le nombre peut ordonner les travaux, et qu'aucune étiquette de catégorie n'en découle

Classification

machine, non validée

Prédiction automatique; un appel candidat d’une seule tête enseignante, pas un consensus.

Les modèles n’ont appliqué aucune catégorie : rien dans la taxonomie ne correspondait à ce travail.
Devis d'étudeAutre devis
Domainenon disponible
GenreEmpirique

Le détail, modèle par modèle et score par score, se trouve en fin de page sous « Comment cette classification a été obtenue ».

En bref

Citations4
Publié2015
Routes d'admission1
Résumé présentoui

Explorer davantage

Même revueInternational Forum of Teaching and StudiesMême sujetNatural Language Processing TechniquesTravaux en français237 207