Neuro-Evolution in Flappy Bird Game

This paper presents an implementation of an Artificial Intelligence (AI) using a Genetic Algorithm (GA) and Deep Neural Network (DNN), with the goal of creating an AI that can win in the game of Flappy Bird. The neural network is a straightforward dense neural network with a single layer of concealment. The application of GA proves to be essential to the achievement of the AI's goals, particularly the mutation of the weight vector. The game starts off with a thousand birds, and a neural network determines whether each bird needs to jump at the beginning of each round. Each bird starts off with its own network already established. The genetic algorithm oversees both the mutation process and the generation of the subsequent population. After a certain number of repetitions, the network, with the help of mutation, eventually figures out how to win the game. The p5.js JavaScript graphics package is also utilized in this work. This was a wonderful and comprehensive introduction to the process of making visuals in a smooth manner on the front end.