Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning
Model-agnostic Measure of Generalization Difficulty
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Increase task input intrinsic dimensionality (e.g., adding perturbations off the image manifold) on ImageNet | ImageNet classification · ImageNet with added perturbations | task difficulty | Helped |
| Increase number of training points | ImageNet task difficulty · ImageNet | task difficulty | Hurt |
| Lower desired error rate | Reacher task difficulty · Reacher | task difficulty | Hurt |
| Use a better architecture (ViT-H/14 vs linear/ResNet) to quantify inductive bias contributed | CIFAR10 and ImageNet classification · CIFAR10 / ImageNet | information content of inductive bias: CIFAR10: 2.250 (Linear) to 3.513 (ViT-H/14); ImageNet: 2.063 to 2.461 | Helped |
| Add Gaussian noise to task inputs (suggested) to increase intrinsic dimension | supervised tasks generally · theory and examples | inductive bias complexity | Helped |
Closest papers by meaning
- Do Neural Networks Need Gradient Descent to Generalize? A Theoretical Study (NeurIPS 2025 · similarity 0.86)
- Position: The No Free Lunch Theorem, Kolmogorov Complexity, and the Role of Inductive Biases in Machine Learning (ICML 2024 · similarity 0.86)
- Fantastic Generalization Measures are Nowhere to be Found (ICLR 2024 · similarity 0.86)
- Towards Generalization beyond Pointwise Learning: A Unified Information-theoretic Perspective (ICML 2024 · similarity 0.86)
- On the Limitations of Fractal Dimension as a Measure of Generalization (NeurIPS 2024 · similarity 0.85)
- Simplicity Bias in Overparameterized Machine Learning (AAAI 2024 · similarity 0.85)
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.