AI RESEARCH

Reparameterization through Coverings and Topological Weight Priors

arXiv CS.LG

ArXi:2604.23804v1 Announce Type: new We generalise the reparameterization trick applied in variational autoencoders (VAEs) letting these have latent spaces of non-trivial topology - i.e. that of base manifolds covered with other ones, on which some technique for RT is available.