multi-stakeholder fairness aware recommendation systems
Keywords: fairness, recommendation systems
TL;DR: This paper introduces the topic of fairness aware recommendation algorithms, using an intersectional definition of fairness, for discussion and further collaboration with the hope to develop it into a publishable paper.
Abstract: This research focuses on multi-stakeholder fairness aware recommenders. That is, recommendation systems that must consider the welfare of multiple stakeholders other than just the end user in the recommendations made. This research focuses on the role that an initial algorithm has on the overall performance of fairness aware re-ranking and how the accuracy-fairness trade-off is affected as accuracy increases. It explores the evaluation of fairness from the perspective of the Black feminist theory of intersectionality. Overall the aim of this research is to inform the design of fairness-aware recommendation systems adopting a definition of fairness that is grounded in the theory of intersectionality.
This paper introduces the topic for discussion and further collaboration with the hope to develop it into a publishable paper.
Submission Category: Machine learning algorithms
Submission Number: 81
Loading