Published January 30, 2022 | Version 1.0.0

FA*IR: A fair Top-k Ranking Algorithm - A Reproducibility Study

Description

Ranked search results are omnipresent in our day to day life and we heavily rely on them. We increasingly depend on algorithms that provide search results from large amounts of data and assist us in decision making. While the inner workings of these algorithms is often well protected by companies, the results that these algorithms return affect decision-making and have impacts in peoples lives. This has lead to concerns about the fairness of ranking in search results with respect to vulnerable groups, e.g. underrepresented groups in the job market. The FA*IR algorithm introduced by Zehlike et al. aims to tackle this problem. In this reproducibility study, we attempt to reproduce the results of the paper where FA*IR is introduced, achieving satisfactory results and providing further evidence of the effectiveness of FA*IR.

Files

ExDDS_Report_Group03.pdf

Files (529.3 kB)

Name Size
md5:c27eac482ee8fd16ce3cf9927bfa20fd
529.3 kB Preview Download