Fractal Graph Contrastive Learning
About
Graph Contrastive Learning (GCL) relies on semantically consistent graph augmentations, but common local perturbations provide limited control over global structural consistency, motivating a more principled global augmentation strategy. We therefore propose Fractal Graph Contrastive Learning (FractalGCL), a theory-motivated framework that constructs a renormalisation-based augmented graph and introduces a fractal-dimension-aware contrastive loss that penalises unreliable positive views and reweights negative-pair repulsion by finite-scale box-counting discrepancies. However, computing these discrepancies introduces substantial overhead, so we derive and justify a Gaussian surrogate that avoids repeated box-counting on renormalised graphs, yielding about a $61\%$ runtime reduction. Experiments show that FractalGCL serves as an effective frozen-pretraining tool on MalNet-Tiny, achieves strong performance on the standard TUDataset benchmarks, and outperforms the next-best method on real-world urban traffic tasks by $4.51$ percentage points in average accuracy. Code is available at https://anonymous.4open.science/r/FractalGCL-0511/.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Graph Classification | PROTEINS | Accuracy75.85 | 1252 | |
| Graph Classification | MUTAG | Accuracy91.71 | 1103 | |
| Graph Classification | NCI1 | Accuracy80.5 | 658 | |
| Graph Classification | D&D | Accuracy80.14 | 146 | |
| Graph Classification | REDDIT-M5K | Accuracy0.5645 | 24 | |
| Graph Classification | MalNet Tiny | Accuracy93.4 | 21 | |
| Graph Classification | Urban road-network subgraphs Chicago, San Francisco, New York (averaged over all cities and tasks) | Accuracy61.76 | 6 |