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

Pretrained machine learning models are known to perpetuate and even amplifyexisting biases in data, which can result in unfair outcomes that ultimatelyimpact user experience. Therefore, it is crucial to understand the mechanismsbehind those prejudicial biases to ensure that model performance does notresult in discriminatory behaviour toward certain groups or populations. Inthis work, we define gender bias as our case study. We quantify biasamplification in pretraining and after fine-tuning on three families ofvision-and-language models. We investigate the connection, if any, between thetwo learning stages, and evaluate how bias amplification reflects on modelperformance. Overall, we find that bias amplification in pretraining and afterfine-tuning are independent. We then examine the effect of continuedpretraining on gender-neutral data, finding that this reduces groupdisparities, i.e., promotes fairness, on VQAv2 and retrieval tasks withoutsignificantly compromising task performance.

Evaluating Bias and Fairness in Gender-Neutral Pretrained Vision-and-Language Models