AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

Module-wise Training of Neural Networks via the Minimizing Movement Scheme

Skander Karkar, Ibrahim Ayed, Emmanuel de Bezenac, patrick gallinari

NeurIPS 2023 · Architectures and efficiency › RNNs and normalization · approach: Differentiable optimization layers · 2 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
add transport regularization (TRGL vs VanGL) (ablation)parallel module-wise training, ResNet-110, 16 modules · STL10test accuracy (%): 73.57 -> 77.22Improved
add transport regularization (TRGL vs VanGL) (ablation)parallel module-wise training, ResNet-110, 4 modules · STL10test accuracy (%): 77.11 -> 79.72Improved
add transport regularization (TRGL vs VanGL) (ablation)parallel, ResNet-101, 4 modules · TinyImageNettest accuracy (%): 63.64 -> 63.71No clear effect
add transport regularization (TRGL vs VanGL) (ablation)parallel, Swin-Tiny, 4 modules · CIFAR100test accuracy (%): 69.34 -> 74.11Improved
add transport regularization (TRGL vs VanGL) (ablation)parallel, 2-7 ResNet · CIFAR10test accuracy (%): 94.01 -> 94.05No clear effect
add transport regularization (ablation)sequential block-wise 10-1 ResNet · CIFAR10, train size 1000last block test accuracy (%): 45.67 -> 51.34Improved
add transport regularization (ablation)sequential block-wise 10-1 ResNet · CIFAR100, train size 50000highest test accuracy (%): 68.74 -> 68.79No clear effect
add transport regularization (ablation)parallel 10-1 ResNet · CIFAR100, train size 5000highest test accuracy (%): 39.45 -> 40.36Improved
use multi-lap sequential training instead of sequential (ablation)TRGL, 10-1 ResNet · CIFAR100, train size 50000highest test accuracy (%): 68.79 -> 69.95Improved
use multi-lap sequential training instead of sequential (ablation)VanGL, 10-1 ResNet · CIFAR100, train size 5000highest test accuracy (%): 36.37 -> 33.68Hurt
add multi-lap sequential training to sequential (ablation)ResNet-101 split into 2 modules · CIFAR100test accuracy (%): 73.31 -> 74.34Improved
add regularization to multi-lap sequential (ablation)ResNet-101 split into 2 modules · CIFAR100test accuracy (%): 74.34 -> 74.78Improved

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