Thompson sampling is a technique to solve multi-armed bandit problems but choosing actions based on the probability they offer the highest expected reward. In this post we learn about this technique by implementing an ad auction!