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

Numerous works use word embedding-based metrics to quantify societal biasesand stereotypes in texts. Recent studies have found that word embeddings cancapture semantic similarity but may be affected by word frequency. In this workwe study the effect of frequency when measuring female vs. male gender biaswith word embedding-based bias quantification methods. We find that Skip-gramwith negative sampling and GloVe tend to detect male bias in high frequencywords, while GloVe tends to return female bias in low frequency words. We showthese behaviors still exist when words are randomly shuffled. This proves thatthe frequency-based effect observed in unshuffled corpora stems from propertiesof the metric rather than from word associations. The effect is spurious andproblematic since bias metrics should depend exclusively on word co-occurrencesand not individual word frequencies. Finally, we compare these results with theones obtained with an alternative metric based on Pointwise Mutual Information.We find that this metric does not show a clear dependence on frequency, eventhough it is slightly skewed towards male bias across all frequencies.

The Undesirable Dependence on Frequency of Gender Bias Metrics Based on Word Embeddings