1009

Our paper “Coupling-based invertible neural networks are universal diffeomorphism approximators” has been accepted at NeurIPS 2020 for oral presentation! (105 orals among the 1900 accepted papers; paper acceptance rate 20.1%).