Research map / Learning theory and algorithms / Kernels and regression theory
Rethinking Influence Functions of Neural Networks in the Over-Parameterized Regime
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 neural-network width from 1×10^4 to 8×10^4 (ablation) | NTK IF approximation · MNIST | Pearson's R: 0.9766 -> 0.9955 | Helped |
| Increase neural-network width from 1×10^4 to 8×10^4 (ablation) | NTK IF approximation · MNIST | Spearman's ρ: 0.9341 -> 0.9728 | Helped |
| Increase neural-network width from 1×10^4 to 8×10^4 (ablation) | NTK IF approximation · CIFAR-10 | Pearson's R: 0.9683 -> 0.9960 | Helped |
| Increase neural-network width from 1×10^4 to 8×10^4 (ablation) | NTK IF approximation · CIFAR-10 | Spearman's ρ: 0.8583 -> 0.9606 | Helped |
| Use stronger regularization for IHVP estimation | IHVP influence-function estimate · Over-parameterized neural networks | estimation accuracy | Helped |
| Estimate influence for a training point from a higher-density data region | IHVP influence-function estimate · MNIST and CIFAR-10 | approximation error | Helped |
Closest papers by meaning
- Towards Robust Influence Functions with Flat Validation Minima (ICML 2025 · similarity 0.87)
- Neural Networks for Learnable and Scalable Influence Estimation of Instruction Fine-Tuning Data (NeurIPS 2025 · similarity 0.86)
- Deep Neural Networks Tend To Extrapolate Predictably (ICLR 2024 · similarity 0.85)
- Rethinking Benign Overfitting in Two-Layer Neural Networks (ICML 2025 · similarity 0.85)
- Sharper Guarantees for Learning Neural Network Classifiers with Gradient Methods (ICLR 2025 · similarity 0.85)
- An Infinite-Width Analysis on the Jacobian-Regularised Training of a Neural Network (ICML 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.