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

The quality of automatic metrics for machine translation has beenincreasingly called into question, especially for high-quality systems. Thispaper demonstrates that, while choice of metric is important, the nature of thereferences is also critical. We study different methods to collect referencesand compare their value in automated evaluation by reporting correlation withhuman evaluation for a variety of systems and metrics. Motivated by the findingthat typical references exhibit poor diversity, concentrating aroundtranslationese language, we develop a paraphrasing task for linguists toperform on existing reference translations, which counteracts this bias. Ourmethod yields higher correlation with human judgment not only for thesubmissions of WMT 2019 English to German, but also for Back-translation andAPE augmented MT output, which have been shown to have low correlation withautomatic metrics using standard references. We demonstrate that ourmethodology improves correlation with all modern evaluation metrics we look at,including embedding-based methods. To complete this picture, we reveal thatmulti-reference BLEU does not improve the correlation for high quality output,and present an alternative multi-reference formulation that is more effective.

BLEU might be Guilty but References are not Innocent