1 paper · 1 filter
Srivaths Ranganathan, Zihuan Diao, Bernardo Cunha +7
Continuously trained ranking models in music recommenders fall into feedback loops where previously consumed items dominate recommendations. This suppresses two distinct content cl…