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 works have found evidence of gender bias in models of machinetranslation and coreference resolution using mostly synthetic diagnosticdatasets. While these quantify bias in a controlled experiment, they often doso on a small scale and consist mostly of artificial, out-of-distributionsentences. In this work, we find grammatical patterns indicating stereotypicaland non-stereotypical gender-role assignments (e.g., female nurses versus maledancers) in corpora from three domains, resulting in a first large-scale genderbias dataset of 108K diverse real-world English sentences. We manually verifythe quality of our corpus and use it to evaluate gender bias in variouscoreference resolution and machine translation models. We find that all testedmodels tend to over-rely on gender stereotypes when presented with naturalinputs, which may be especially harmful when deployed in commercial systems.Finally, we show that our dataset lends itself to finetuning a coreferenceresolution model, finding it mitigates bias on a held out set. Our dataset andmodels are publicly available at www.github.com/SLAB-NLP/BUG. We hope they willspur future research into gender bias evaluation mitigation techniques inrealistic settings.

Collecting a Large-Scale Gender Bias Dataset for Coreference Resolution and Machine Translation