Tfrom

Abstract

At present, most research on the fairness of recommender systems is conducted either from the perspective of customers or from the perspective of product (or service) providers. However, such a practice ignores the fact that when fairness is guaranteed to one side, the fairness and rights of the other side are likely to reduce. In this paper, we consider recommendation scenarios from the perspective of two sides (customers and providers). From the perspective of providers, we consider the fairness of the providers' exposure in recommender system. For customers, we consider the fairness of the reduced quality of recommendation results due to the introduction of fairness measures. We theoretically analyzed the relationship between recommendation quality, customers fairness, and provider fairness, and design a two-sided fairness-aware recommendation model (TFROM) for both customers and providers. Specifically, we design two versions of TFROM for offline and online recommendation. The effectiveness of the model is verified on three real-world data sets. The experimental results show that TFROM provides better two-sided fairness while still maintaining a higher level of personalization than the baseline algorithms.

P-MMF: Provider Max-min Fairness Re-ranking in Recommender System

“…As for the evaluation metrics, the performances of the models were evaluated from three aspects: user-side preference, providerside fairness, and the trade-off between them. As for the userside preference, following the practices in [47], we utilized the NDCG@K, which is defined as the ratio between the sum of position-based user-item scores [47] in the original ranking list L 𝐾 (𝑢 𝑡 ) and that in the re-ranked list L 𝐹 𝐾 (𝑢 𝑡 ):…”

“…Rahmani et al [38] studied the trade-offs between the user and producer fairness in Point-of-Interest recommendations. TFROM [47] and CPFair [32] formulated the trade-off as a knapsack problem and a relaxed linear programming problem, respectively. However, they used greedy-based algorithms in online scenarios, which only improves the proportion fairness of providers.…”

“…As for the hyper-parameters in all models, the learning rate was tuned among [1𝑒 − 2, 1𝑒 − 4]/𝑇 1/2 and the momentum coefficient 𝛼 was tuned among [0.2, 0.6]. For the maximum resources (i.e., the weights) 𝜸 , following the practices in [47], we set 𝜸 based on the number of items provided by the providers: 𝜸 𝑝 = 𝐾𝑇 𝜂|I 𝑝 | |I|, where 𝜂 is the factor controlling the richness of resources. In all the experiments, we set 𝜂 = 1 + 1/|P |.…”

“…Existing provider fair recommendation models either consider the PF of providers [8,32,35,47] or use heuristics to guarantee the utilities of the worse-off providers [8,35]. However, these heuristic methods do not directly consider the MMF and lack theoretical guarantees when adapting to online scenarios.…”

Two-sided fairness-aware recommendation method under many-to-many relation schema

“…We extend previous works 5,6 to propose a two-sided fairness-aware recommendation model TwoFair (Two-Sided Fairness-Aware Recommendation Model Based on Fairness Allocation Under Many-to-many Relation Schema), which can be used for two-sided fairness-aware recommendations when the relationship pattern between providers and items is many-to-many, to improve the two-sided fairness of recommendation results while maintaining the recommendation quality as much as possible.…”