Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A Survey on Trustworthy Recommender Systems

Yingqiang Ge, Shuchang Liu, Zuohui Fu, Juntao Tan, Zelong Li, Shuyuan Xu, Yunqi Li, Yikun Xian, Yongfeng Zhang

TL;DR

This survey defines Trustworthy Recommender Systems (TRS) as a multi-perspective framework encompassing explainability, fairness, privacy, robustness, and controllability. It synthesizes over 400 works to map the methods, definitions, and evaluations across these dimensions, and elucidates how these perspectives interact and trade off with one another. By detailing foundational RS concepts, explainable generation and evaluation, fairness notions, privacy threats and protections, robustness against attacks, and user-centric controllability, the paper provides a holistic blueprint for designing trust-driven recommender systems. The work highlights practical implications, including the cost of trustworthiness, cross-perspective compatibility, and evaluation gaps, and outlines open challenges and opportunities to guide future research toward more trustworthy and human-centered RS.

Abstract

Recommender systems (RS), serving at the forefront of Human-centered AI, are widely deployed in almost every corner of the web and facilitate the human decision-making process. However, despite their enormous capabilities and potential, RS may also lead to undesired effects on users, items, producers, platforms, or even the society at large, such as compromised user trust due to non-transparency, unfair treatment of different consumers, or producers, privacy concerns due to extensive use of user's private data for personalization, just to name a few. All of these create an urgent need for Trustworthy Recommender Systems (TRS) so as to mitigate or avoid such adverse impacts and risks. In this survey, we will introduce techniques related to trustworthy recommendation, including but not limited to explainable recommendation, fairness in recommendation, privacy-aware recommendation, robustness in recommendation, user-controllable recommendation, as well as the relationship between these different perspectives in terms of trustworthy recommendation. Through this survey, we hope to deliver readers with a comprehensive view of the research area and raise attention to the community about the importance, existing research achievements, and future research directions on trustworthy recommendation.

A Survey on Trustworthy Recommender Systems

TL;DR

This survey defines Trustworthy Recommender Systems (TRS) as a multi-perspective framework encompassing explainability, fairness, privacy, robustness, and controllability. It synthesizes over 400 works to map the methods, definitions, and evaluations across these dimensions, and elucidates how these perspectives interact and trade off with one another. By detailing foundational RS concepts, explainable generation and evaluation, fairness notions, privacy threats and protections, robustness against attacks, and user-centric controllability, the paper provides a holistic blueprint for designing trust-driven recommender systems. The work highlights practical implications, including the cost of trustworthiness, cross-perspective compatibility, and evaluation gaps, and outlines open challenges and opportunities to guide future research toward more trustworthy and human-centered RS.

Abstract

Recommender systems (RS), serving at the forefront of Human-centered AI, are widely deployed in almost every corner of the web and facilitate the human decision-making process. However, despite their enormous capabilities and potential, RS may also lead to undesired effects on users, items, producers, platforms, or even the society at large, such as compromised user trust due to non-transparency, unfair treatment of different consumers, or producers, privacy concerns due to extensive use of user's private data for personalization, just to name a few. All of these create an urgent need for Trustworthy Recommender Systems (TRS) so as to mitigate or avoid such adverse impacts and risks. In this survey, we will introduce techniques related to trustworthy recommendation, including but not limited to explainable recommendation, fairness in recommendation, privacy-aware recommendation, robustness in recommendation, user-controllable recommendation, as well as the relationship between these different perspectives in terms of trustworthy recommendation. Through this survey, we hope to deliver readers with a comprehensive view of the research area and raise attention to the community about the importance, existing research achievements, and future research directions on trustworthy recommendation.
Paper Structure (39 sections, 5 figures, 1 table)

This paper contains 39 sections, 5 figures, 1 table.

Table of Contents

  1. Introduction
  2. Trustworthy Perspectives & Their Relationships in Recommendation
  3. Explainability with other Perspectives
  4. Fairness with other Perspectives
  5. Privacy with other Perspectives
  6. Robustness with other Perspectives
  7. Controllability with other Perspectives
  8. Recommender System Basics
  9. Input and Output of RS
  10. Representative Recommendation Algorithms
  11. Explainability
  12. Overview of Explainable Recommendation
  13. Different Styles of Explanations in Recommender Systems
  14. Explanation Generation in Recommender Systems
  15. Explanations Evaluation in Recommender Systems
  16. ...and 24 more sections

Figures (5)

  • Figure 1: Trustworthiness is a multi-perspective concept in AI and recommender systems.
  • Figure 2: Typical inputs of a recommender system.
  • Figure 4: Different dimensions to define fairness in recommendation.
  • Figure 5: Three types of fair recommendation methods in different parts of the recommendation pipeline.
  • Figure 6: Ownership types (left), privacy threats (middle), and protection techniques (right).