Our paper “A Geometric Perspective on Variational Autoencoders” is accepted to NeurIPS 2022!