This project compares the parameters used for a generative algorithm with the features learned by a Variational Autoncoder (VAE) trained to reconstruct the generated ouput objects. The final aim is to compare the parameter space with the latent space learned by the ML model.