In my previous post, I conducted a fairly thorough walkthrough of the game of blackjack, reinforcement learning, Q learning, and various analyses. In this post, I’ll introduce the concept of Deep Reinforcement Learning (or, Deep Q Learning).Why Deep Q Learning?In traditional frameworks for Q Learning, we store state-action pairs and their associated Q value in memory. In the game of blackjack, this is a tractable solution, only if we exclude card count entirely. To review, let’s take an arbit...