FlapAI Bird: Training an Agent to Play Flappy Bird Using Reinforcement Learning Techniques
This is an incremental application of existing reinforcement learning methods to a specific game, with no broader implications.
The authors tackled the problem of training an agent to play Flappy Bird using reinforcement learning, finding that SARSA and Q-Learning outperformed a baseline with scores regularly over 1400 and a highest score of 2069.
Reinforcement learning is one of the most popular approaches for automated game playing. This method allows an agent to estimate the expected utility of its state in order to make optimal actions in an unknown environment. We seek to apply reinforcement learning algorithms to the game Flappy Bird. We implement SARSA and Q-Learning with some modifications such as $ε$-greedy policy, discretization and backward updates. We find that SARSA and Q-Learning outperform the baseline, regularly achieving scores of 1400+, with the highest in-game score of 2069.