Reinforcement Learning for Slate-based Recommender Systems: A Tractable Decomposition and Practical Methodology

To read later

link