Permutation-Invariant Deep Reinforcement Learning for Danmaku Games

Takuto ITOI

Bachelor Thesis, 31 Jan, 2022

We propose a new model that is able to handle permutation-invariant input data with low computational power, which shows better performance in playing Danmaku games compared to the SOTA methods. We also created a Gym environment for Danmaku games that eases further researches.