Taming Filter Bubbles in Recommender Systems: Simulation and Diversification

Project: Research

View graph of relations

Description

Personalized recommender systems have become significantly important in modern society to reduce information/choice overload. Despite being helpful and valuable, recommender systems are criticized as a primary cause of information "filter bubble" or "echo chamber", where users are trapped in an algorithm-created bubble without the opportunity to view diverse contents. This leads to severely negative consequences such as fake news proliferation and political polarization.Although recent studies have managed to analyze and burst the filter bubble in recommender systems, there are still twofold challenges: 1) lacking a dynamic environment to analyze the evolution of filter bubbles in the long run; 2) lacking general solutions to address the accuracy­diversity dilemma when breaking filter bubbles.In this project, we will conduct systematic investigations on analyzing and breaking filter bubbles in recommender systems. Our plan involves two major tasks: 1) analyze filter bubbles in a dynamic environment, and 2) diversify recommended items against filter bubbles.

Detail(s)

Project number7020082
Grant typeSIRG
StatusActive
Effective start/end date1/04/23 → …