TFROM: A Two-sided Fairness-Aware Recommendation Model for Both Customers and Providers

However, recommender At present, most research on the fairness of recommender systems systems can also bring unfavorable consequences, such is conducted either from the perspective of customers or from the as they may narrow the customers' vision [1], or superior items perspective of product(or service) providers. However, such a practice will receive increased attention so as to become dominant [27], ignores the fact that when fairness is guaranteed to one side, while inferior items will be relegated to a lower position, which the fairness and rights of the other side are likely to reduce. In becomes an extremely vicious circle. As a possible unfavorable consequence, this paper, we consider recommendation scenarios from the perspective the unfairness in recommender systems in different aspects, of two sides(customers and providers). From the perspective such as racial/gender stereotypes [22], social polarization of providers, we consider the fairness of the providers' exposure [12], position bias [27], has been a well-studied research topic. in recommender system. For customers, we consider the fairness Problem Statement. Despite the different mechanisms which of the reduced quality of recommendation results due to the have been implemented to ensure the fairness of recommendations, introduction of fairness measures. We theoretically analyzed the these studies only consider the utility of one type of stakeholder relationship between recommendation quality, customers fairness, in business and try to eliminate unfairness among their members.