Recommender systems are gaining increasing and critical impacts on human and society since a growing number of users use them for information seeking and decision making . Just like users have personalized preferences on items, users’ demands for fairness are also personalized in many scenarios . To this end, weintroduce a framework for achieving counterfactually fair recommendationsthrough adversary learning by generating feature-independent user embeddings for recommendation . The framework allows recommender systems to achieve personalized fairness for users while also covering non-personalizedsituations. Experiments on two real-world datasets with shallow and deeprecommendation algorithms show that our method can generate fairerrecommendations for users with a desirable recommendation performance. The method can be used in real-time data with a desired recommendation performance, according to the authors . The method is published in The New Scientist
Author(s) : Yunqi Li, Hanxiong Chen, Shuyuan Xu, Yingqiang Ge, Yongfeng ZhangLinks : PDF - Abstract
Code :
Keywords : users - personalized - fairness - method - recommendation -