Ensuring provider fairness in recommender systems across coarse- and fine-grained groups

The goal of provider fairness in recommender systems is to ensure equity by suggesting products from diverse providers or provider groups. When group fairness is among the goals of a system, coarse groups are frequently used, since there are typically few provider groups (e.g., two genders, or three/four age groups) and the number of items per group is large. Practically speaking, having fewer groups makes it easier for a platform to oversee how equity is distributed among them. Nevertheless, there are sensitive attributes, such as the age or the geographic provenance of the providers, that can be characterized at a fine granularity (e.g., one might group providers at the country level, instead of the continent one), which increases the number of groups and decreases the number of items per group. This study reveals that state-of-the-art models often fail to adequately recommend fine-grained provider groups when only coarse-grained groups are considered. This oversight can result in a fairness approach that, while adequate for broader demographic groups, neglects the needs of smaller subgroups. To address this disparity, we introduce CONFIGRE (COarse aNd FIne GRained Equity), an approach designed to balance equity across both coarse and fine-grained provider groups. Our methodology ensures that fairness is not only maintained at a broad demographic level but is also extended to more precisely defined groups, offering a more nuanced and comprehensive equity management in recommender systems. • We show the limitations of using coarse-grained groups in recommender systems for ensuring provider fairness. • We present CONFIGRE, a novel approach to balance equity across both coarse and fine-grained provider groups. • We validate, via experiments on real-world data, that CONFIGRE extends fairness beyond broad demographic levels to more precisely defined provider groups.