A Deep Q-Network with Experience Optimization (DQN-EO) for Atari’s Space Invaders
Since Deep Q-Learning was introduced in 2013, a lot of academia and industry researchers, tried to use it to solve their problems, in different fields. However, Deep Q-Learning does not consist of a unified method for solving certain problems. In contrary, every problem requires specific settings and parameters. In this paper, we propose a Deep Q-Network with Experience Optimization, for Atari’s “Space Invaders” environment. Training and Testing results show that, while using our proposed network, there is a strong correlation between the key action (shoot) and scoring.
KeywordsArtificial intelligence Deep Reinforcement Learning Atari Space Invaders Experience Optimization
- 2.OpenAI Gym Toolkit. https://gym.openai.com/
- 3.Mnih, V., Kavukcuoglu, K., Silver, D., Graves, A., Antonoglou, I., Wierstra, D., Riedmiller, M.: Playing Atari with Deep Reinforcement Learning. arXiv, 19 December 2013. https://arxiv.org/pdf/1312.5602v1.pdf
- 4.Mnih, V., Kavukcuoglu, K., Silver, D., Rusu, A., Veness, J., Bellemare, M.G., Graves, A., Riedmiller, M., Fidjeland, A.K., Ostrovski, G., Petersen, S., Beattie, C., Sadik, A., Antonoglou, I., King, H., Kumaran, D., Wierstra, D., Legg, S., Hassabis, D.: Human-level control through deep reinforcement learning. Nature. 518, 529–33 (2015). https://doi.org/10.1038/nature14236CrossRefGoogle Scholar