flâneur

Spotify Reinforcement Learning Recommendation System

sophieyanzhao.github.io · 2,519 words · saved by 1 readers

Team members: Feng Qian, Sophie Zhao, Yizhou Wang Recommendation system can be a vital competitive edge for service providers such as Spotify, who mainly grows business through user subscriptions. Accurate recommendations help improve user experience and strengthen customer loyalty. Traditional recommendation methods include modeling user-item interaction with supervised learning such...

Team members: Feng Qian, Sophie Zhao, Yizhou Wang Recommendation system can be a vital competitive edge for service providers such as Spotify, who mainly grows business through user subscriptions. Accurate recommendations help improve user experience and strengthen customer loyalty. Traditional recommendation methods include modeling user-item interaction with supervised learning such as classification, memory-based content-filtering from user history and many more. These ideas overlook the dependency across consecutive time steps. Inspired by the progress of reinforcement learning in other do

related reading