Research map / Optimization / Implicit bias and generalization
On global convergence of ResNets: From finite to infinite width using linear parameterization
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 |
|---|---|---|---|
| Use fixed hidden layers in residual blocks instead of training them (ablation) | RKHS-NODE · MNIST | accuracy: 1% accuracy drop | Hurt |
| Use fixed hidden layers in residual blocks instead of training them (ablation) | RKHS-NODE · MNIST, pretrained A and B | accuracy: 0.5% accuracy | Hurt |
| Pretrain A and B (ablation) | RKHS-NODE · MNIST test set | accuracy: up to 98% -> more than 99% | Helped |
| Increase input channels from 4 to 8 or 32 (ablation) | RKHS-NODE · MNIST | empirical risk convergence | Helped |
| Increase NODE-block depth from 1 layer to 10 or 20 layers (ablation) | RKHS-NODE · CIFAR10 | accuracy | Helped |
| Increase NODE-block depth from 10 to 20 layers (ablation) | RKHS-NODE · CIFAR10 | accuracy | No clear effect |
| Use fixed hidden layers in residual blocks instead of training them (ablation) | RKHS-NODE · CIFAR10 test set | accuracy | No clear effect |
Closest papers by meaning
- On the Global Convergence of Training Deep Linear ResNets (ICLR 2020 · similarity 0.91)
- Generalization of Scaled Deep ResNets in the Mean-Field Regime (ICLR 2024 · similarity 0.91)
- A Mean Field Analysis Of Deep ResNet And Beyond: Towards Provably Optimization Via Overparameterization From Depth (ICML 2020 · similarity 0.90)
- Are deep ResNets provably better than linear predictors? (NeurIPS 2019 · similarity 0.90)
- Approximation theory for 1-Lipschitz ResNets (NeurIPS 2025 · similarity 0.88)
- Block Coordinate Descent for Neural Networks Provably Finds Global Minima (NeurIPS 2025 · similarity 0.88)
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.