CPFair: Personalized Consumer and Producer Fairness Re-ranking for Recommender Systems

Figure 2 illustrates the distribution of research on various aspects of fairness in recommender systems.111The We may observe a division/split in the research on fairness-aware recommendation algorithms, with around 49.1% of papers concentrating on consumer fairness and somewhat fewer on producer fairness (41.8%). Few studies (less than 10%) address consumer and producer fairness concerns simultaneously. However, the underlying user-item interaction coexists on (and can impact) both sides of beneficiary stakeholders. For example, there may be disparities in the consumption of item groups (defined by protected attributes) between active and inactive users in specific domains.