AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

Path-conditioned training: a principled way to rescale ReLU neural networks

Arthur Lebeurrier, Titouan Vayer, Rémi Gribonval

ICML 2026 · Architectures and efficiency › RNNs and normalization · approach: Theory of depth and width · 0 citations (OpenAlex)

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
rescale ReLU network at initialization with PathCondSGD, Kaiming-initialized MLP · CIFAR-10 MLPepochs to 99% training accuracy: up to 1.5x fasterImproved
rescale at initialization with PathCondSGD, CIFAR-NV · CIFAR-10best test accuracy: 77 -> 80Improved
replace Equinormalization rescaling with PathCondSGD MLP · CIFAR-10 MLPepochs to 99% training accuracyImproved
add PathCond with increasing network depth (more hidden layers) (ablation)SGD MLP · CIFAR-10 MLPtraining speedupMixed
increase learning rate (ablation)PathCond · CIFAR-10speedup over baselineMixed
increase compression factor (more unbalanced autoencoder) (ablation)PathCond · MNIST autoencoderfinal training lossImproved
use PathCond on autoencoder at all compression factorsSGD · MNIST autoencoderfinal training lossImproved
vary layer width irregularity and init variance (ablation)PathCond · random ReLU netslog rescaling magnitudeMixed

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