Hidden or Inferred: Fair Learning-To-Rank with Unknown Demographics

Oluseun Olulana, Kathleen Cachel, Fabricio Murai, Elke Rundensteiner

Abstract: As learning-to-rank models are increasingly deployed for decision-making in areas with profound life implications, the FairML community has been developing fair learning-to-rank (LTR) models. These models rely on the availability of sen- sitive demographic features such as race or sex. However, in practice, regulatory obstacles and privacy concerns pro- tect this data from collection and use. As a result, practi- tioners may either need to promote fairness despite the ab- sence of these features or turn to demographic inference tools to attempt to infer them. Given that these tools are fallible, this paper aims to further understand how errors in demo- graphic inference impact the fairness performance of popu- lar fair LTR strategies. In which cases would it be better to keep such demographic attributes hidden from models ver- sus infer them? We examine a spectrum of fair LTR strate- gies ranging from fair LTR with and without demographic features hidden versus inferred to fairness-unaware LTR fol- lowed by fair re-ranking. We conduct a controlled empirical investigation modeling different levels of inference errors by systematically perturbing the inferred sensitive attribute. We also perform three case studies with real-world datasets and popular open-source inference methods. Our findings reveal that as inference noise grows, LTR-based methods that incor- porate fairness considerations into the learning process may increase bias. In contrast, fair re-ranking strategies are more robust to inference errors. All source code, data, and experi- mental artifacts of our experimental study are available here: https://github.com/sewen007/hoiltr.git