Chained NEAT implementation for better adapting neural networks applied to Super Mario World

Abstract

Con uso de redes neuronales encadenadas basadas en NEAT se crea un prototipo de un agente, juega el videojuego "Super Mario World" con posibilidad de adaptarse a varios niveles.

Using chained neural networks based on NEAT, we developed a prototype capable of playing "Super Mario World" with a better capability of adapting to different levels