PD3: Better Low-Resource Cross-Lingual Transfer By Combining Direct Transfer and Annotation Projection

Steffen Eger, Andreas Rücklé, Iryna Gurevych


Abstract
We consider unsupervised cross-lingual transfer on two tasks, viz., sentence-level argumentation mining and standard POS tagging. We combine direct transfer using bilingual embeddings with annotation projection, which projects labels across unlabeled parallel data. We do so by either merging respective source and target language datasets or alternatively by using multi-task learning. Our combination strategy considerably improves upon both direct transfer and projection with few available parallel sentences, the most realistic scenario for many low-resource target languages.
Anthology ID:
W18-5216
Volume:
Proceedings of the 5th Workshop on Argument Mining
Month:
November
Year:
2018
Address:
Brussels, Belgium
Editors:
Noam Slonim, Ranit Aharonov
Venue:
ArgMining
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
131–143
Language:
URL:
https://aclanthology.org/W18-5216
DOI:
10.18653/v1/W18-5216
Bibkey:
Cite (ACL):
Steffen Eger, Andreas Rücklé, and Iryna Gurevych. 2018. PD3: Better Low-Resource Cross-Lingual Transfer By Combining Direct Transfer and Annotation Projection. In Proceedings of the 5th Workshop on Argument Mining, pages 131–143, Brussels, Belgium. Association for Computational Linguistics.
Cite (Informal):
PD3: Better Low-Resource Cross-Lingual Transfer By Combining Direct Transfer and Annotation Projection (Eger et al., ArgMining 2018)
Copy Citation:
PDF:
https://aclanthology.org/W18-5216.pdf
Code
 UKPLab/emnlp2018-argmin-workshop-pd3