Research map / Learning theory and algorithms / Kernels and regression theory
Generalization for multiclass classification with overparameterized linear models
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 |
|---|---|---|---|
| Scale the number of classes as k = c_k n^t while using minimum-norm interpolation with bi-level weighted Gaussian features | Multiclass classification · Bi-level ensemble, 1-sparse noiseless labels | asymptotic misclassification probability | Helped |
| Use multiclass rather than binary labels with the same total number of training points | Minimum-norm interpolating classifier · Bi-level ensemble | survival-to-contamination ratio: decreases by a factor of √k | Hurt |
| Increase the number of training samples k-fold to match the number of positive examples per class to the binary case | Multiclass minimum-norm interpolating classifier · Bi-level ensemble | asymptotic generalization | Helped |
| Use multiclass classification in the regime q+r>1 where the corresponding regression problem does not generalize | Minimum-norm interpolating classifier · Bi-level ensemble | asymptotic misclassification probability | Helped |
Closest papers by meaning
- Precise asymptotic generalization for multiclass classification with overparameterized linear models (NeurIPS 2023 · similarity 0.93)
- Theoretical Insights Into Multiclass Classification: A High-dimensional Asymptotic View (NeurIPS 2020 · similarity 0.90)
- Polynomial-Time Robust Multiclass Linear Classification under Gaussian Marginals (NeurIPS 2026 · similarity 0.87)
- Trading off Consistency and Dimensionality of Convex Surrogates for Multiclass Classification (NeurIPS 2024 · similarity 0.85)
- Provable weak-to-strong generalization via benign overfitting (ICLR 2025 · similarity 0.85)
- Multiclass Loss Geometry Matters for Generalization of Gradient Descent in Separable Classification (NeurIPS 2025 · 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.