Q-learning grid lab

Train an agent to cross a grid world from rewards alone, and watch its values spread out from the goal. This tool needs JavaScript, so turn it on to use it. Everything you need to answer the exam question is explained in the text above.