NeurIPS 2019
Sun Dec 8th through Sat the 14th, 2019 at Vancouver Convention Center
Paper ID:6458
Title:Graph Transformer Networks

The paper proposes a graph transformer network for multirelational graphs. Thee network learns to transform the input graph into useful meta-path graphs for generating new graphs and for learning node representations. While the reviewers point out some weaknesses, we overall agree that this is an interesting piece of work addressing an important problem.