NeurIPS 2019
Sun Dec 8th through Sat the 14th, 2019 at Vancouver Convention Center
Paper ID:5712
Title:Variational Graph Recurrent Neural Networks

The proposed method is a sound combination of existing methods (GRNN [20] (deterministic) and GVAE [13] (no smoothness in time)), providing impressive performance gain for dynamic/new link prediction. Interpretability is stated but NOT supported by any. The authors should correct the paper for the camera ready so that all statements are supported.