Research map / Learning theory and algorithms / Kernels and regression theory
Generalization Bounds using Data-Dependent Fractal Dimensions
Benjamin Dupuis, George Deligiannidis, Umut Simsekli
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Replace Euclidean trajectory distance with the loss-based data-dependent pseudo-metric | FCN-5 · CHD | Spearman's rank correlation (ρ): 0.77 -> 0.87 | Improved |
| Replace Euclidean trajectory distance with the loss-based data-dependent pseudo-metric | FCN-7 · CHD | Spearman's rank correlation (ρ): 0.40 -> 0.77 | Improved |
| Replace Euclidean trajectory distance with the loss-based data-dependent pseudo-metric | FCN-5 · MNIST | Spearman's rank correlation (ρ): 0.62 -> 0.73 | Improved |
| Replace Euclidean trajectory distance with the loss-based data-dependent pseudo-metric | FCN-7 · MNIST | Spearman's rank correlation (ρ): 0.80 -> 0.89 | Improved |
| Replace Euclidean trajectory distance with the loss-based data-dependent pseudo-metric | AlexNet · CIFAR-10 | Spearman's rank correlation (ρ): 0.86 -> 0.93 | Improved |
| Estimate the data-dependent pseudo-metric from a random subset rather than the full training set | PH dimension computation · MNIST and CHD | Relative error in PH dimension: 40% | No clear effect |
Closest papers by meaning
- On the Limitations of Fractal Dimension as a Measure of Generalization (NeurIPS 2024 · similarity 0.93)
- Topological Generalization Bounds for Discrete-Time Stochastic Optimization Algorithms (NeurIPS 2024 · similarity 0.89)
- Generalizability of Neural Networks Minimizing Empirical Risk Based on Expressive Power (ICLR 2025 · similarity 0.88)
- Generalization at the Edge of Stability (NeurIPS 2026 · similarity 0.87)
- Towards Generalization beyond Pointwise Learning: A Unified Information-theoretic Perspective (ICML 2024 · similarity 0.87)
- Fantastic Generalization Measures are Nowhere to be Found (ICLR 2024 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
