Projektbeschreibung
Bessere neuronale maschinelle Übersetzung für seltene Sprachen
Da genaue und schnell verfügbare Informationen heutzutage unerlässlich sind, brauchen Journalistinnen und Journalisten quasi täglich entsprechende Möglichkeiten, um schnell korrekte Übersetzungen aus Sprachen zu erhalten, die seltener vorkommen. Die Technologie der neuronalen maschinellen Übersetzung entwickelt sich zwar in Rekordtempo, doch noch immer gibt es für die meisten denkbaren Sprachpaare keine verwendbaren Übersetzungen, da Daten und Korpora von Paralleltexten fehlen. Das EU-finanzierte Projekt GoURMET will nun Belastbarkeit und Anwendbarkeit von neuronaler maschineller Übersetzung für seltene Sprachpaare und -gruppen verbessern. Maschinelle Übersetzungen sollen von Menschen korrigiert und internationale Nachrichtenmedien im Hinblick auf seltene Sprachpaare verfolgt werden, um global relevante Inhalte zu erstellen.
Ziel
Machine translation (MT) is an increasingly important technology for supporting communication in a globalised world. MT technology has gradually increased over the last ten years, but recent advances in neural machine translation (NMT), have resulted in significant interest in industry and have lead to very rapid adoption of the new paradigm (eg. Google, Facebook, UN, World International Patent Office). Although these models have shown significant advances in state-of-the-art performance they are data intensive and require parallel corpora of many millions of human translated sentences for training. Neural Machine translation is currently not able to deliver usable translations for the vast majority of language pairs in the world. This is especially problematic for our user partners, the BBC and DW who need access to fast and accurate translation for languages with very few resources.
The aim of GoURMET is to significantly improve the robustness and applicability of neural machine translation for low-resource language pairs and domains.
GoURMET has five objectives:
- Development of a high-quality machine translation for under-resourced language pairs and domains;
- Adaptable to new and emerging languages and domains;
- Development of tools for analysts and journalists;
- Sustainable, maintainable platform and services;
- Dissemination and communication of project results to stakeholders and user group.
The project will focus on two use cases:
- Global content creation - managing content creation in several languages efficiently by providing machine translations for correction by humans;
- Media monitoring for low resource language pairs - tools to address the challenge of international news monitoring problem.
The outputs of the project will be field-tested at partners BBC and DW, and the platform will be further validated through innovation intensives such as the BBC NewsHack.
Wissenschaftliches Gebiet
Schlüsselbegriffe
Programm/Programme
Aufforderung zur Vorschlagseinreichung
Andere Projekte für diesen Aufruf anzeigenUnterauftrag
H2020-ICT-2018-2
Finanzierungsplan
RIA - Research and Innovation actionKoordinator
EH8 9YL Edinburgh
Vereinigtes Königreich