FLAPPY BIRD AI USING REINFORCEMENT LEARNING

Abstract - Reinforcement learning plays a crucial role in solving problems where multiple solutions exist. Flappy Bird AI applies deep reinforcement learning to train an agent that learns to play Flappy Bird without human intervention. Despite having no prior knowledge of the bird or pipes, the AI analyzes game states and scores to develop an optimal strategy. A Convolutional Neural Network (CNN) processes visual input, while Q-learning with Deep Q-Networks (DQN) helps the AI make smart decisions. By balancing exploration and exploitation, superhuman performance will be attained in navigating obstacles. Enhancements like experience replay and target networks will improve learning efficiency. Flappy Bird AI will demonstrate the potential of reinforcement learning in autonomous gaming and intelligent decision-making systems. Key Words: Flappy Bird AI, Reinforcement Learning, Deep-Q-Network, convolutional Neural Network, Q-Learning, Autonomous Gaming AI,Game AI.