Incentivized Bandit Learning with Self-Reinforcing User Preferences

               Tianchen Zhou 1 Jia Liu 1 Chaosheng Dong 2 Jingyuan Deng 2

             Abstract               accumulates more positive feedbacks. For example, on a
    In this paper, we investigate a new multi-armed     movie rental website, current customers tend to have more
    bandit (MAB) online learning model that consid-     interest in Movie A that has 500 positive reviews, compared
    ers real-world phenomena in man ...