Recent advancements in deep generative modeling, and especially Generative Adversarial Networks (GANs), propose to replace the original input data with features from a pre-trained network to allow a more stable training procedure and working in a low-data regime. The recent ProjectedGAN paper exhibited great results in generating new images when there is very low data available (e.g., images of Pokemons). In this project, we will investigate ways to improve this approach and experimenting with Variational Autoencoders (VAEs) instead of GANs. The project may involve different data types, such as images, 3D data or tabular data. The students will build their own generative model and try to improve the generative capabilities over a baseline algorithm, gaining hands-on experience in deep learning, unsupervised learning and deep generative models.