论文信息 - Nintendo Super Smash Bros. Melee: An "Untouchable" Agent

Nintendo Super Smash Bros. Melee: An "Untouchable" Agent

Nintendo's Super Smash Bros. Melee fighting game can be emulated on modern hardware allowing us to inspect internal memory states, such as character positions. We created an AI that avoids being hit by training using these internal memory states and outputting controller button presses. After training on a month's worth of Melee matches, our best agent learned to avoid the toughest AI built into the game for a full minute 74.6% of the time.

[1] Tom Schaul,et al. Dueling Network Architectures for Deep Reinforcement Learning , 2015, ICML.

[2] Alex Graves,et al. Asynchronous Methods for Deep Reinforcement Learning , 2016, ICML.

[3] Richard S. Sutton,et al. Introduction to Reinforcement Learning , 1998 .

[4] Yuval Tassa,et al. Continuous control with deep reinforcement learning , 2015, ICLR.

[5] David Silver,et al. Deep Reinforcement Learning with Double Q-Learning , 2015, AAAI.

[6] Joshua B. Tenenbaum,et al. Beating the World's Best at Super Smash Bros. with Deep Reinforcement Learning , 2017, ArXiv.

[7] Shane Legg,et al. Human-level control through deep reinforcement learning , 2015, Nature.