This repository contains code for generating and predicting the frames of Conway’s Game of Life using deep learning. Conway’s Game of Life (CGOL) is a cellular automaton that evolves according to a simple set of rules on a two-dimensional grid. The goal is to learn a model that can generate the next states of the grid given the current state or extrapolate multiple steps into the future, thereby testing the limits of computational irreducibility.
The model used is a convolutional neural network (CNN) that takes advantage of the local patterns and symmetries in the game. The CNN is trained on synthetic data generated by applying the rules of the game to random initial configurations. The repository also includes some examples of the model’s predictions and a notebook for visualizing and evaluating the results.