Oct 8, 2015 · We introduce the unwarranted associations (UA) framework, a principled methodology for the discovery of unfair, discriminatory, or offensive ...

Aug 16, 2016 · Abstract. In a world where traditional notions of privacy are increasingly challenged by the myriad companies.

We introduce the unwarranted associations (UA) framework, a principled methodology for the discovery of unfair, discriminatory, or offensive user treatment in ...

FairTest [23] is a prominent example that uncovers unwarranted associations in predictive models. For instance, if a job recommendation system ...

We describe FairTest, a testing toolkit that detects unwarranted associations between an algorithm's outputs (e.g., prices or labels) and user subpopulations, ...

Foremost challenge is to even detect these unwarranted associations. Page 9. FairTest: a testing suite for data-driven apps. Data-driven application. User ...

3 FairTest Overview. Fig.1(a) shows the FairTest architecture. At a high level, the data-driven application – the object of FairTest's investigations – takes ...

FairTest is a testing toolkit that detects unwarranted associations between an algorithm's outputs and user subpopulations and ranks them by their strength ...

FairTest enables developers or auditing entities to discover and test for unwarranted associations between an algorithm's outputs and certain user ...

We describe FairTest, a testing toolkit that detects unwarranted associations between an algorithm's outputs (e.g., prices or labels) and user subpopulations, ...