推荐系统中公平性问题研究现状与展望  

Research Status and Prospect of Fairness in Recommendation System

作　　者：李文俊 LI Wen-jun(Suzhou Vocational Institute of Industrial Technology,Suzhou 215000,China)

机构地区：[1]苏州工业职业技术学院,苏州215000

出　　处：《价值工程》2023年第5期166-168,共3页Value Engineering

基　　金：江苏省自然科学基金青年项目《基于时间感知的多样性小众推荐算法技术研究》,项目编号:BK20180209;江苏省高校自然科学基金项目《基于时序感知的小众推荐算法研究》,项目编号:18KJD520004;苏州工业职业技术学院科研启动经费项目“《基于时间感知的推荐技术研究》”研究成果,项目编号:2019kyqd018。

摘　　要：随着数字经济的快速发展,推荐系统已经深入在线生活的众多领域。随着在决策类应用中嵌入推荐功能,推荐系统中的公平性问题成为了一个十分值得重视的概念,不公平的原因通常来自于偏见,不公平的推荐结果可能会损害用户或者内容提供者的权益。近年来,大量研究开始聚焦于向用户和内容提供者提供满足各自公平性需求的推荐算法模型。本文对近期的相关工作,对推荐场景下的公平性问题的来源、定义和算法模型进行了总结,并在最后简述了相关领域仍需解决的问题和未来的发展方向。With the rapid development of the digital economy,recommendation systems have penetrated into many areas of online life.With the embedding of recommendation function in decision-making applications,the fairness problem in the recommendation system has become a very important concept.The reason for unfairness usually comes from bias,and the unfair recommendation results may damage the rights and interests of users or content providers.In recent years,a large number of studies have begun to focus on providing users and content providers with recommendation algorithm models that meet their fairness needs.This paper summarizes the source,definition and algorithm model of the fairness problem in the recommendation scenario based on the recent related work,and briefly describes the problems that still need to be solved in the relevant fields and the future development direction.

关 键 词：推荐算法 公平性 偏见 

分 类 号：TP391.3[自动化与计算机技术—计算机应用技术]