An image of several runs during training.

Deep Reinforcement Learning and Transfer Learning with FlappyBird

Reinforcement learning’s growth in popularity in recent years is partly due to its ability to play some video games with a level of mastery that no human can reach. Transfer learning is popular in the field of deep learning, and using pre-trained models on certain tasks speeds up training time and increases performance significantly. In this project we aimed to apply transfer learning to the popular video game FlappyBird and analyze its performance compared to traditional reinforcement learning algorithms.

Please read our report here.

Cedrick Argueta
Computer Science Student