Entry
How Algorithmic Confounding in Recommendation Systems Increases Homogeneity and Decreases Utility
Allison J. B. Chaney, Brandon M. Stewart, Barbara E. Engelhardt
Simulation studies show that data collected under algorithmic recommendations is confounded — homogenizes user behavior without increasing utility.
·algorithmic confounding ·recommender feedback ·homogeneity ·simulation