Human Interaction with Recommendation Systems

Abstract

Many recommendation algorithms rely on user data to generate recommendations. However, these recommendations also affect the data obtained from future users. This work aims to understand the effects of this dynamic interaction. We propose a simple model where users with heterogeneous preferences arrive over time. Based on this model, we prove that naive estimators, i.e. those which ignore this feedback loop, are not consistent. We show that consistent estimators are efficient in the presence of myopic agents. Our results are validated using extensive simulations.

Accepted to AISTATS 2018

Cited by in corpus (11)

• Towards Fair Personalization by Avoiding Feedback Loops
• Scalable Realistic Recommendation Datasets through Fractal Expansions
• Do Offline Metrics Predict Online Performance in Recommender Systems?
• Human self-determination within algorithmic sociotechnical systems
• Bandit Learning with Delayed Impact of Actions
• Correcting the User Feedback-Loop Bias for Recommendation Systems
• T-RECS: A Simulation Tool to Study the Societal Impact of Recommender Systems
• Scaling Up Collaborative Filtering Data Sets through Randomized Fractal Expansions
• Bayesian Exploration with Heterogeneous Agents
• A Bayesian Choice Model for Eliminating Feedback Loops
• Bandit based centralized matching in two-sided markets for peer to peer lending