REVEAL@RecSys2020: Workshop on Bandit and Reinforcement Learning from User Interactions
State-of-the-art recommender systems are notoriously hard to design and improve upon, due to their interactive and dynamic nature. In particular, they involve a multi-step decision-making...