Open AccessProceedings Article
Weisfeiler and Lehman Go Topological: Message Passing Simplicial Networks
Cristian Bodnar,Fabrizio Frasca,Yu Guang Wang,Yu Guang Wang,Yu Guang Wang,Nina Otter,Guido Montúfar,Pietro Liò,Michael M. Bronstein +8 more
- pp 1026-1037
Reads0
Chats0
TLDR
In this paper, the authors proposed Message Passing Simplicial Networks (MPSNs), a class of models that perform message passing on simplicial complexes (SCs) and introduced a Simplicial Weisfeiler-Lehman (SWL) colouring procedure for distinguishing non-isomorphic SCs.Abstract:
The pairwise interaction paradigm of graph machine learning has predominantly governed the modelling of relational systems. However, graphs alone cannot capture the multi-level interactions present in many complex systems and the expressive power of such schemes was proven to be limited. To overcome these limitations, we propose Message Passing Simplicial Networks (MPSNs), a class of models that perform message passing on simplicial complexes (SCs). To theoretically analyse the expressivity of our model we introduce a Simplicial Weisfeiler-Lehman (SWL) colouring procedure for distinguishing non-isomorphic SCs. We relate the power of SWL to the problem of distinguishing non-isomorphic graphs and show that SWL and MPSNs are strictly more powerful than the WL test and not less powerful than the 3-WL test. We deepen the analysis by comparing our model with traditional graph neural networks (GNNs) with ReLU activations in terms of the number of linear regions of the functions they can represent. We empirically support our theoretical claims by showing that MPSNs can distinguish challenging strongly regular graphs for which GNNs fail and, when equipped with orientation equivariant layers, they can improve classification accuracy in oriented SCs compared to a GNN baseline.read more
Citations
More filters
Posted Content
Hodgelets: Localized Spectral Representations of Flows on Simplicial Complexes.
TL;DR: In this article, the authors developed wavelet representations for edge-flows on simplicial complexes, using ideas rooted in combinatorial Hodge theory and spectral graph wavelets, and characterized the representational quality of their Hodgelets for edge flows in terms of frame bounds.
Posted Content
Topological Graph Neural Networks.
TL;DR: TOGL as mentioned in this paper incorporates global topological information of a graph using persistent homology, which can be easily integrated into any type of GNN and is strictly more expressive in terms of the Weisfeiler-Lehman test of isomorphism.
Posted Content
From Stars to Subgraphs: Uplifting Any GNN with Local Structure Awareness.
TL;DR: GNN-AK as discussed by the authors is a general framework to uplift any MPNN to be more expressive, with limited scalability overhead and greatly improved practical performance by extending local aggregation in MPNNs from star patterns to general subgraph patterns (e.g., k-egonets).