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

In budget-constrained settings aimed at mitigating unfairness, like lawenforcement, it is essential to prioritize the sources of unfairness beforetaking measures to mitigate them in the real world. Unlike previous works,which only serve as a caution against possible discrimination and de-bias dataafter data generation, this work provides a toolkit to mitigate unfairnessduring data generation, given by the Unfair Edge Prioritization algorithm, inaddition to de-biasing data after generation, given by the DiscriminationRemoval algorithm. We assume that a non-parametric Markovian causal modelrepresentative of the data generation procedure is given. The edges emanatingfrom the sensitive nodes in the causal graph, such as race, are assumed to bethe sources of unfairness. We first quantify Edge Flow in any edge X -> Y,which is the belief of observing a specific value of Y due to the influence ofa specific value of X along X -> Y. We then quantify Edge Unfairness byformulating a non-parametric model in terms of edge flows. We then prove thatcumulative unfairness towards sensitive groups in a decision, like race in abail decision, is non-existent when edge unfairness is absent. We prove thisresult for the non-trivial non-parametric model setting when the cumulativeunfairness cannot be expressed in terms of edge unfairness. We then measure thePotential to mitigate the Cumulative Unfairness when edge unfairness isdecreased. Based on these measurements, we propose the Unfair EdgePrioritization algorithm that can then be used by policymakers. We also proposethe Discrimination Removal Procedure that de-biases a data distribution byeliminating optimization constraints that grow exponentially in the number ofsensitive attributes and values taken by them. Extensive experiments validatethe theorem and specifications used for quantifying the above measures.

A Causal Approach for Unfair Edge Prioritization and Discrimination Removal