1 paper
Nicolas Legendre, Khanh Dao Duc, Nina Miolane
Recent advances in variational autoencoders (VAEs) have enabled learning latent manifolds as compact Lie groups, such as SO(d). Since this approach assumes that data lies on a su…