495-project-Qbert Using rainbow to play Qbert - which feature in Rainbow has the biggest effect on learning Youtube: https://youtu.be/kVLe_XmBGc4 Vanilla Q-learning deepQ_keras_Qbert_v5.ipynb Rainbow with modifications Test_all.ipynb Results Rainbow Q Reward Rainbow without prioritized memory replay Q Reward Rainbow without Dueling DQN (turned out to be the most important) Q Reward Rainbow without Lossy Net Q Reward C51 to C5 distributional learning Q Reward