AI Paper Atlas

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

ICLR 2026 · Architectures and efficiency › RNNs and normalization · approach: Representation geometry and collapse

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 changedWhereMeasuredEffect
plot test error against spectral-complexity norm instead of epochsCNN, ResNet, ViT · MNIST, CIFAR10, CIFAR100learning-curve scalingImproved
combine gamma_1 and gamma_2 to predict the end-of-training exponent (ablation)deep nets · CIFAR10 CNN weight decay 1e-3exponent: gamma_pred 0.163 vs gamma_meas 0.212Improved
add moderate weight decay (ablation)CNN, ResNet, ViT · CIFAR10gamma_predNo clear effect
use SGD instead of Adam (ablation)CNN · CIFAR10, CIFAR100gamma_pred: CIFAR10 0.202 vs 0.225No clear effect
use SGD instead of Adam (ablation)CNN · CIFAR10, CIFAR100dynamical exponentsMixed
use alternative norms instead of spectral complexity (ablation)deep nets · image classificationgamma_pred vs gamma_measHurt
use MSE loss instead of logistic loss (ablation)perceptron · teacher-studentscaling lawsHurt

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy