Geometry-Aware Edge Pooling for Graph Neural Networks
BP2-STS
- Limbeck, Katharina ORCID Helmholtz Munich; Technical University of Munich
- Mezrag, Lydia Université de Montréal; Mila - Quebec AI Institute
- Wolf, Guy ORCID Université de Montréal; Mila - Quebec AI Institute
- Rieck, Bastian ORCID Helmholtz Munich; Université de Fribourg
- 2026
Published in:
- 39th Conference on Neural Information Processing Systems (NeurIPS 2025). - 2026
English
Graph Neural Networks (GNNs) have shown significant success for graph-based tasks. Motivated by the prevalence of large datasets in real-world applications, pooling layers are crucial components of GNNs. By reducing the size of input graphs, pooling enables faster training and potentially better generalisation. However, existing pooling operations often optimise for the learning task at the expense of discarding fundamental graph structures, thus reducing interpretability. This leads to unreliable performance across dataset types, downstream tasks and pooling ratios. Addressing these concerns, we propose novel graph pooling layers for structure-aware pooling via edge collapses. Our methods leverage diffusion geometry anditeratively reduce a graph's size while preserving both its metric structure and its structural diversity. We guide pooling using magnitude, an isometry-invariant diversity measure, which permits us to control the fidelity of the pooling process. Further, we use the spread of a metric space as a faster and more stable alternative ensuring computational efficiency. Empirical results demonstrate that our methods (i) achieve top performance compared to alternative pooling layers across a range of diverse graph classification tasks, (ii) preserve key spectral properties of the input graphs, and (iii) retain high accuracy across varying pooling ratios.
- Faculty
- Faculté des sciences et de médecine
- Department
- Département d'Informatique
- Language
-
- English
- Classification
- Computer science and technology
- License
-
License undefined
- Open access status
- green
- Persistent URL
- https://folia.unifr.ch/unifr/documents/336331
Statistics
Document views: 2
File downloads:
- neurips-2025-geometry-aware-edge-pooling-for-graph-neural-networks-paper-conference: 0