Research map / Architectures and efficiency / RNNs and normalization
Implicit bias produces neural scaling laws in learning curves, from perceptrons to deep networks
Francesco D'Amico, Dario Bocchi, Matteo Negri
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| plot test error against spectral-complexity norm instead of epochs | CNN, ResNet, ViT · MNIST, CIFAR10, CIFAR100 | learning-curve scaling | Improved |
| combine gamma_1 and gamma_2 to predict the end-of-training exponent (ablation) | deep nets · CIFAR10 CNN weight decay 1e-3 | exponent: gamma_pred 0.163 vs gamma_meas 0.212 | Improved |
| add moderate weight decay (ablation) | CNN, ResNet, ViT · CIFAR10 | gamma_pred | No clear effect |
| use SGD instead of Adam (ablation) | CNN · CIFAR10, CIFAR100 | gamma_pred: CIFAR10 0.202 vs 0.225 | No clear effect |
| use SGD instead of Adam (ablation) | CNN · CIFAR10, CIFAR100 | dynamical exponents | Mixed |
| use alternative norms instead of spectral complexity (ablation) | deep nets · image classification | gamma_pred vs gamma_meas | Hurt |
| use MSE loss instead of logistic loss (ablation) | perceptron · teacher-student | scaling laws | Hurt |
Closest papers by meaning
- A Dynamical Model of Neural Scaling Laws (ICML 2024 · similarity 0.93)
- Scaling Laws in Linear Regression: Compute, Parameters, and Data (NeurIPS 2024 · similarity 0.90)
- Analyzing Neural Scaling Laws in Two-Layer Networks with Power-Law Data Spectra (ICLR 2025 · similarity 0.90)
- Navigating Scaling Laws: Compute Optimality in Adaptive Model Training (ICML 2024 · similarity 0.89)
- A Hitchhiker's Guide to Scaling Law Estimation (ICML 2025 · similarity 0.88)
- On the Invariance and Generality of Neural Scaling Laws (NeurIPS 2026 · 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.
