Existence conditions for hidden feedback loops in online recommender systems

TL;DR

This paper investigates the conditions under which hidden feedback loops occur in online recommender systems, showing how noise and user interest resets influence their existence and impact.

Contribution

It provides theoretical analysis and experimental validation of how noise and interest resets affect feedback loops in multi-armed bandit recommender systems.

Findings

Unbiased additive noise does not prevent feedback loops.

Interest resets can limit feedback loop effects.

Experimental results confirm theoretical predictions.

Abstract

We explore a hidden feedback loops effect in online recommender systems. Feedback loops result in degradation of online multi-armed bandit (MAB) recommendations to a small subset and loss of coverage and novelty. We study how uncertainty and noise in user interests influence the existence of feedback loops. First, we show that an unbiased additive random noise in user interests does not prevent a feedback loop. Second, we demonstrate that a non-zero probability of resetting user interests is sufficient to limit the feedback loop and estimate the size of the effect. Our experiments confirm the theoretical findings in a simulated environment for four bandit algorithms.