Mario RL

• An RL agent plays Super Mario
• Algorithm used is Double Deep Q Network
• Our model has shown the improvement with episodes.
• For good results it has to be trained with much higher number of episodes.

Necessary Libraries

• Super Mario Gym

   pip install gym-super-mario-bros==7.3.0

• nes-wrappers, Open AI gym, PyTorch, numpy, basic python libraries

References

• Deep Reinforcement Learning with Double Q-learning (https://arxiv.org/pdf/1509.06461)
• Pytorch Tutorials

This project is done under the guidance of Vision And Language Group @IIT Roorkee