Cascading Bandits for Large-Scale Recommendation Problems

Shi Zong, Hao Ni, Kenny Sung +3

Most recommender systems recommend a list of items. The user examines the list, from the first item to the last, and often chooses the first attractive item and does not examine th…