Impact of Domain-Adapted Multilingual Neural Machine Translation in the Medical Domain

Veröffentlichungen: Beitrag in BuchBeitrag in KonferenzbandPeer Reviewed

Abstract

Multilingual Neural Machine Translation (MNMT) models leverage many language pairs during training to improve translation quality for low-resource languages by transferring knowledge from high-resource languages. We study the quality of a domain-adapted MNMT model in the medical domain for English-Romanian with automatic metrics and a human error typology annotation which includes terminology-specific error categories. We compare the out-of-domain MNMT with the in-domain adapted MNMT. The in-domain MNMT model outperforms the out-of-domain MNMT in all measured automatic metrics and produces fewer terminology errors.
OriginalspracheEnglisch
TitelProceedings of the 44th Translating and the Computer (TC44) conference
Redakteure*innenJoss Moorkens, Vilelmini Sosoni
ErscheinungsortGeneva
Herausgeber (Verlag)Editions Tradulex
Seiten27-38
ISBN (elektronisch)978-2-9701733-0-4
PublikationsstatusVeröffentlicht - Sep. 2023
VeranstaltungTranslating and the Computer - European Convention Center Luxembourg (ECCL), Luxembourg, Luxemburg
Dauer: 24 Nov. 202225 Nov. 2022
Konferenznummer: 44
https://asling.org/tc44/

Konferenz

KonferenzTranslating and the Computer
KurztitelTC44
Land/GebietLuxemburg
OrtLuxembourg
Zeitraum24/11/2225/11/22
Internetadresse

ÖFOS 2012

  • 602051 Translationswissenschaft

Fingerprint

Untersuchen Sie die Forschungsthemen von „Impact of Domain-Adapted Multilingual Neural Machine Translation in the Medical Domain“. Zusammen bilden sie einen einzigartigen Fingerprint.

Zitationsweisen