Few-Shot and Zero-Shot Learning for Historical Text Normalization

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

Documents

Historical text normalization often relies on small training datasets. Recent work has shown that multi-task learning can lead to significant improvements by exploiting synergies with related datasets, but there has been no systematic study of different multi-task learning architectures. This paper evaluates 63 multi-task learning configurations for sequence-to-sequence-based historical text normalization across ten datasets from eight languages, using autoencoding, grapheme-to-phoneme mapping, and lemmatization as auxiliary tasks. We observe consistent, significant improvements across languages when training data for the target task is limited, but minimal or no improvements when training data is abundant. We also show that zero-shot learning outperforms the simple, but relatively strong, identity baseline.
Original languageEnglish
Title of host publicationProceedings of the 2nd Workshop on Deep Learning Approaches for Low-Resource NLP (DeepLo 2019)
PublisherAssociation for Computational Linguistics
Publication date2019
Pages104-114
DOIs
Publication statusPublished - 2019
Event2nd Workshop on Deep Learning Approaches for Low-Resource NLP (DeepLo) - Hong Kong, China
Duration: 3 Nov 20193 Nov 2019

Workshop

Workshop2nd Workshop on Deep Learning Approaches for Low-Resource NLP (DeepLo)
LandChina
ByHong Kong
Periode03/11/201903/11/2019

Number of downloads are based on statistics from Google Scholar and www.ku.dk


No data available

ID: 239617207