AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

Padé Activation Units: End-to-end Learning of Flexible Activation Functions in Deep Networks

Alejandro Molina, Patrick Schramowski, Kristian Kersting

ICLR 2020 · Architectures and efficiency › RNNs and normalization · approach: Theory of depth and width · 30 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
replace fixed activations with learnable PAUsVGG-8 on MNIST · MNISTtest accuracy: 99.17 ±0.10 -> 99.30 ±0.05No clear effect
replace fixed activations with learnable PAUsLeNet on Fashion-MNIST · Fashion-MNISTtest accuracy: 89.86 ±0.32 -> 90.33 ±0.15Improved
add coefficient noise during training (ablation)PAU on MNIST · MNIST, LeNettest accuracy: 99.21 \pm0.04 -> 99.26 \pm0.11Improved
add coefficient noise during training (ablation)PAU on Fashion-MNIST · Fashion-MNIST, LeNettest accuracy: 90.33 \pm0.15 -> 90.20 \pm0.11Hurt
add coefficient noise during training (ablation)PAU on CIFAR-10 · CIFAR-10, MobileNetV2test accuracy: 94.57 \pm0.21 -> 94.82 \pm0.21Improved
add coefficient noise during training (ablation)PAU on CIFAR-10 · CIFAR-10, ResNet101test accuracy: 95.16 \pm0.13 -> 95.34 \pm0.13Improved
add coefficient noise during training (ablation)PAU on CIFAR-10 · CIFAR-10, DenseNet121test accuracy: 95.03 \pm0.07 -> 95.27 \pm0.10Improved
replace fixed activations with learnable PAUsMobileNetV2 · ImageNettop-1 accuracy: 71.24 -> 71.35Improved
replace fixed activations with learnable PAUsMobileNetV2 · ImageNettop-5 accuracy: 89.95 -> 89.85Hurt
prune convolutional layers of PAU networks (ablation)ResNet101 with PAU · CIFAR-10accuracyNo clear effect
prune the PAU ResNet101 subnetwork by 30% (ablation)ResNet101 with PAU · CIFAR-10accuracyImproved
replace fixed activations with learnable activationsVGG-8 and MobileNetV2 · CIFAR-10test accuracyImproved

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