English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Pretrained multilingual models (PMMs) enable zero-shot learning viacross-lingual transfer, performing best for languages seen during pretraining.While methods exist to improve performance for unseen languages, they havealmost exclusively been evaluated using amounts of raw text only available fora small fraction of the world's languages. In this paper, we evaluate theperformance of existing methods to adapt PMMs to new languages using a resourceavailable for over 1600 languages: the New Testament. This is challenging fortwo reasons: (1) the small corpus size, and (2) the narrow domain. Whileperformance drops for all approaches, we surprisingly still see gains of up to$17.69\%$ accuracy for part-of-speech tagging and $6.29$ F1 for NER on averageover all languages as compared to XLM-R. Another unexpected finding is thatcontinued pretraining, the simplest approach, performs best. Finally, weperform a case study to disentangle the effects of domain and size and to shedlight on the influence of the finetuning source language.

How to Adapt Your Pretrained Multilingual Model to 1600 Languages