用分形结构增强图对比学习，提升模型泛化能力。
Fractal Graph Contrastive Learning
- 基于重标度构建全局一致的图增广，避免局部扰动失真。
- 在真实交通数据上比最优方法高4.51%准确率，性能显著。
- 引入高斯近似加速计算，运行时间减少61%，适合实际应用。
图对比学习（GCL）依赖语义一致的图增广，但常见的局部扰动难以控制全局结构一致性，因此我们提出分形图对比学习（FractalGCL），一种基于理论的框架。该框架通过重标度构造增广图，并设计分形维数感知的对比损失，惩罚不可靠的正样本视图，同时根据有限尺度盒计数差异重加权负样本排斥。然而，盒计数计算开销大，我们推导并验证了一个高斯近似，避免在重标度图上重复进行盒计数，实现约61%的运行时间降低。实验表明，FractalGCL在MalNet-Tiny上作为冻结预训练工具表现有效，在标准TUDataset基准上取得强性能，并在真实城市交通任务中平均准确率领先次优方法4.51个百分点。代码已公开。
原文摘要 · Abstract (English)
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/.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。