Recently, reinforcement learning (RL) and especially deep RL has achieved significant success in a variety of domains including board games such as Go, video games such as classic Atari games and online games such as DoTa 2, Capture-the-Flag, StarCraft II and others.

Simpler versions of RL techniques, notably bandit algorithms, have long been applied to personalization and recommendation systems. In this talk, I explore the history of reinforcement learning applied to personalization and recommendations as well as the state-of-the-art advances in deep RL techniques in this domain.