DQN with model-based exploration: efficient learning on environments with sparse rewards

Re-assign community