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 Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Does the grammatical gender of a language interfere when measuring thesemantic gender information captured by its word embeddings? A number ofanomalous gender bias measurements in the embeddings of gendered languagessuggest this possibility. We demonstrate that word embeddings learn theassociation between a noun and its grammatical gender in grammatically genderedlanguages, which can skew social gender bias measurements. Consequently, wordembedding post-processing methods are introduced to quantify, disentangle, andevaluate grammatical gender signals. The evaluation is performed on fivegendered languages from the Germanic, Romance, and Slavic branches of theIndo-European language family. Our method reduces the strength of grammaticalgender signals, which is measured in terms of effect size (Cohen's d), by asignificant average of d = 1.3 for French, German, and Italian, and d = 0.56for Polish and Spanish. Once grammatical gender is disentangled, theassociation between over 90% of 10,000 inanimate nouns and their assignedgrammatical gender weakens, and cross-lingual bias results from the WordEmbedding Association Test (WEAT) become more congruent with country-levelimplicit bias measurements. The results further suggest that disentanglinggrammatical gender signals from word embeddings may lead to improvement insemantic machine learning tasks.

Measuring Gender Bias in Word Embeddings of Gendered Languages Requires Disentangling Grammatical Gender Signals