Asymptotic Analysis of Empirical Risk Minimization on Entry-wise i.i.d. Heavy-Tailed Data ERR2F5FXC3
| URI | m2://res/ERR2F5FXC3 |
|---|---|
| Kind | preprint |
| Creators | Kaito Takanami, Takashi Takahashi, Yoshiyuki Kabashima |
| Date | 2026-10-06 |
| Subjects | arxiv cond-mat.dis-nn, arxiv cs.LG, arxiv math.ST, arxiv stat.ML, arxiv stat.TH |
| Scope | out |
Expressions
e1: version, preprint
| Date | 2026-10-06 |
|---|---|
| Ids | arXiv 2610.07637 |
Items in this resource
The mentions of the inventory that name this resource. Each name links to the card of its inventory row.
| Item | Kind | Where | Source |
|---|---|---|---|
| Heavy-tail universality for regularized ERM | statement | theorem 5 | arxiv |
| Generalization error of ERM estimator | statement | proposition 2 | arxiv |
| Typical prediction error and tail reliability with many training examples | statement | theorem 6 | arxiv |
| Coefficient error of the Bayes ridge estimator | statement | theorem 7 | arxiv |
| Scalar reduction of the sample problem | statement | proposition 10 | arxiv |
| Campbell formula | statement | proposition 9 | arxiv |
| Gaussian universality for regularized ERM | statement | proposition 4 | arxiv |
| Coordinate-wise versus elliptical heavy tails | statement | theorem 8 | arxiv |
| Statistics of the learned parameters | statement | theorem 3 | arxiv |
Catalogue records
- arxiv: 2610.07637 (work)