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

Increasingly, software is making autonomous decisions in case of criminalsentencing, approving credit cards, hiring employees, and so on. Some of thesedecisions show bias and adversely affect certain social groups (e.g. thosedefined by sex, race, age, marital status). Many prior works on bias mitigationtake the following form: change the data or learners in multiple ways, then seeif any of that improves fairness. Perhaps a better approach is to postulateroot causes of bias and then applying some resolution strategy. This paperpostulates that the root causes of bias are the prior decisions that affect-(a) what data was selected and (b) the labels assigned to those examples. OurFair-SMOTE algorithm removes biased labels; and rebalances internaldistributions such that based on sensitive attribute, examples are equal inboth positive and negative classes. On testing, it was seen that this methodwas just as effective at reducing bias as prior approaches. Further, modelsgenerated via Fair-SMOTE achieve higher performance (measured in terms ofrecall and F1) than other state-of-the-art fairness improvement algorithms. Tothe best of our knowledge, measured in terms of number of analyzed learners anddatasets, this study is one of the largest studies on bias mitigation yetpresented in the literature.

Bias in Machine Learning Software: Why? How? What to do?