Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Higher-order Clustering and Pooling for Graph Neural Networks

About

Graph Neural Networks achieve state-of-the-art performance on a plethora of graph classification tasks, especially due to pooling operators, which aggregate learned node embeddings hierarchically into a final graph representation. However, they are not only questioned by recent work showing on par performance with random pooling, but also ignore completely higher-order connectivity patterns. To tackle this issue, we propose HoscPool, a clustering-based graph pooling operator that captures higher-order information hierarchically, leading to richer graph representations. In fact, we learn a probabilistic cluster assignment matrix end-to-end by minimising relaxed formulations of motif spectral clustering in our objective function, and we then extend it to a pooling operator. We evaluate HoscPool on graph classification tasks and its clustering component on graphs with ground-truth community structure, achieving best performance. Lastly, we provide a deep empirical analysis of pooling operators' inner functioning.

Alexandre Duval, Fragkiskos Malliaros• 2022

Related benchmarks

TaskDatasetResultRank
Graph ClassificationPROTEINS
Accuracy75
994
Graph ClassificationMUTAG
Accuracy84
862
Graph ClassificationNCI1
Accuracy78
501
Graph ClassificationCOLLAB
Accuracy73
422
Graph ClassificationENZYMES
Accuracy36
318
Graph ClassificationDD
Accuracy79
273
Node ClusteringCora
Accuracy38.17
133
Node ClusteringCiteseer
NMI20
130
Graph ClassificationMolHIV
ROC AUC76
88
Graph ClassificationREDDIT-B
Accuracy91
84
Showing 10 of 28 rows

Other info

Follow for update