Cracking Open the Black Box: What Observations Can Tell Us About Reinforcement Learning Agents

Arnaud Dethise, Marco Canini, Srikanth Kandula