Solutions to preference manipulation in recommender systems require knowledge of meta-preferences

Hal Ashton, Matija Franklin

Iterative machine learning algorithms used to power recommender systems often change people's preferences by trying to learn them. Further a recommender can better predict what a u…