Capsule Graph Neural Network

ICLR 2019 Zhang XinyiLihui Chen

The high-quality node embeddings learned from the Graph Neural Networks (GNNs) have been applied to a wide range of node-based applications and some of them have achieved state-of-the-art (SOTA) performance. However, when applying node embeddings learned from GNNs to generate graph embeddings, the scalar node representation may not suffice to preserve the node/graph properties efficiently, resulting in sub-optimal graph embeddings... (read more)

PDF Abstract

Evaluation Results from the Paper


TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK COMPARE
Graph Classification COLLAB CapsGNN Accuracy 79.62% # 7
Graph Classification D&D CapsGNN Accuracy 75.38% # 20
Graph Classification ENZYMES CapsGNN Accuracy 54.67% # 12
Graph Classification IMDb-B CapsGNN Accuracy 73.10% # 11
Graph Classification IMDb-M CapsGNN Accuracy 50.27% # 10
Graph Classification MUTAG CapsGNN Accuracy 86.67% # 22
Graph Classification NCI1 CapsGNN Accuracy 78.35% # 16
Graph Classification PROTEINS CapsGNN Accuracy 76.28% # 22
Graph Classification RE-M12K CapsGNN Accuracy 46.62% # 4
Graph Classification RE-M5K CapsGNN Accuracy 52.88% # 2