Multilingual projection for parsing truly low-resource languages

Zeljko Agic, Anders Trærup Johannsen, Barbara Plank, Hector Martinez Alonso, Natalie Schluter, Anders Søgaard

Publikation: Artikel i tidsskrift og konference artikel i tidsskriftTidsskriftartikelForskningpeer review

Abstract

We propose a novel approach to cross-lingual part-of-speech tagging and dependency parsing for truly low-resource languages. Our annotation projection-based approach yields tagging and parsing models for over 100 languages. All that is needed are freely available parallel texts, and taggers and parsers for resource-rich languages. The empirical evaluation across 30 test languages shows that our method consistently provides top-level accuracies, close to established upper bounds, and outperforms several competitive baselines.
OriginalsprogEngelsk
TidsskriftTransactions of the Association for Computational Linguistics
Vol/bind4
Sider (fra-til)301-312
ISSN2307-387X
StatusUdgivet - jul. 2016
Udgivet eksterntJa

Emneord

  • Cross-lingual part-of-speech tagging
  • Dependency parsing
  • Low-resource languages
  • Annotation projection
  • Parallel texts

Fingeraftryk

Dyk ned i forskningsemnerne om 'Multilingual projection for parsing truly low-resource languages'. Sammen danner de et unikt fingeraftryk.

Citationsformater