Rademacher Meets Colors: More Expressivity, but at What Cost ?
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866909872762650624 |
|---|---|
| author | Carrasco, Martin Netto, Caio F. Deberaldini Martirosyan, Vahan A. Mehrab, Aneeqa Okoyomon, Ehimare Graziani, Caterina |
| author_facet | Carrasco, Martin Netto, Caio F. Deberaldini Martirosyan, Vahan A. Mehrab, Aneeqa Okoyomon, Ehimare Graziani, Caterina |
| contents | The expressive power of graph neural networks (GNNs) is typically understood through their correspondence with graph isomorphism tests such as the Weisfeiler-Leman (WL) hierarchy. While more expressive GNNs can distinguish a richer set of graphs, they are also observed to suffer from higher generalization error. This work provides a theoretical explanation for this trade-off by linking expressivity and generalization through the lens of coloring algorithms. Specifically, we show that the number of equivalence classes induced by WL colorings directly bounds the GNNs Rademacher complexity -- a key data-dependent measure of generalization. Our analysis reveals that greater expressivity leads to higher complexity and thus weaker generalization guarantees. Furthermore, we prove that the Rademacher complexity is stable under perturbations in the color counts across different samples, ensuring robustness to sampling variability across datasets. Importantly, our framework is not restricted to message-passing GNNs or 1-WL, but extends to arbitrary GNN architectures and expressivity measures that partition graphs into equivalence classes. These results unify the study of expressivity and generalization in GNNs, providing a principled understanding of why increasing expressive power often comes at the cost of generalization. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_10101 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Rademacher Meets Colors: More Expressivity, but at What Cost ? Carrasco, Martin Netto, Caio F. Deberaldini Martirosyan, Vahan A. Mehrab, Aneeqa Okoyomon, Ehimare Graziani, Caterina Machine Learning The expressive power of graph neural networks (GNNs) is typically understood through their correspondence with graph isomorphism tests such as the Weisfeiler-Leman (WL) hierarchy. While more expressive GNNs can distinguish a richer set of graphs, they are also observed to suffer from higher generalization error. This work provides a theoretical explanation for this trade-off by linking expressivity and generalization through the lens of coloring algorithms. Specifically, we show that the number of equivalence classes induced by WL colorings directly bounds the GNNs Rademacher complexity -- a key data-dependent measure of generalization. Our analysis reveals that greater expressivity leads to higher complexity and thus weaker generalization guarantees. Furthermore, we prove that the Rademacher complexity is stable under perturbations in the color counts across different samples, ensuring robustness to sampling variability across datasets. Importantly, our framework is not restricted to message-passing GNNs or 1-WL, but extends to arbitrary GNN architectures and expressivity measures that partition graphs into equivalence classes. These results unify the study of expressivity and generalization in GNNs, providing a principled understanding of why increasing expressive power often comes at the cost of generalization. |
| title | Rademacher Meets Colors: More Expressivity, but at What Cost ? |
| topic | Machine Learning |
| url | https://arxiv.org/abs/2510.10101 |