Hierarchical Graph Representation Learning with Differentiable Pooling

Recently, graph neural networks (GNNs) have revolutionized the field of graph representation learning through effectively learned node embeddings, and achieved state-of-the-art results in tasks such as node classification and link prediction. However, current GNN methods are inherently flat and do not learn hierarchical representations of graphs---a limitation that is especially problematic for the task of graph classification, where the goal is to predict the label associated with an entire graph... (read more)

PDF Abstract NeurIPS 2018 PDF NeurIPS 2018 Abstract
TASK DATASET MODEL METRIC NAME METRIC VALUE GLOBAL RANK RESULT BENCHMARK
Graph Classification COLLAB GNN (DiffPool) Accuracy 75.48% # 19
Graph Classification D&D S2V (with 2 DiffPool) Accuracy 82.07% # 5
Graph Classification D&D GNN (DiffPool) Accuracy 80.64% # 10
Graph Classification ENZYMES S2V (with 2 DiffPool) Accuracy 63.33% # 11
Graph Classification ENZYMES GNN (DiffPool) Accuracy 62.53% # 12
Graph Classification PROTEINS GNN (DiffPool) Accuracy 76.25% # 36
Graph Classification REDDIT-MULTI-12K GNN (DiffPool) Accuracy 47.08 # 1

Methods used in the Paper


METHOD TYPE
🤖 No Methods Found Help the community by adding them if they're not listed; e.g. Deep Residual Learning for Image Recognition uses ResNet