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

Recent complementary strands of research have shown that leveraginginformation on the data source through encoding their properties intoembeddings can lead to performance increase when training a single model onheterogeneous data sources. However, it remains unclear in which situationsthese dataset embeddings are most effective, because they are used in a largevariety of settings, languages and tasks. Furthermore, it is usually assumedthat gold information on the data source is available, and that the test datais from a distribution seen during training. In this work, we compare theeffect of dataset embeddings in mono-lingual settings, multi-lingual settings,and with predicted data source label in a zero-shot setting. We evaluate onthree morphosyntactic tasks: morphological tagging, lemmatization, anddependency parsing, and use 104 datasets, 66 languages, and two differentdataset grouping strategies. Performance increases are highest when thedatasets are of the same language, and we know from which distribution thetest-instance is drawn. In contrast, for setups where the data is from anunseen distribution, performance increase vanishes.

On the Effectiveness of Dataset Embeddings in Mono-lingual,Multi-lingual and Zero-shot Conditions