AI Paper Atlas

Research map / Adversarial robustness and security / Certified and empirical robustness

Absum: Simple Regularization Method for Reducing Structural Sensitivity of Convolutional Neural Networks

Sekitoshi Kanai, Yasutoshi Ida, Yasuhiro Fujiwara, Masanori Yamada, Shuichi Adachi

AAAI 2020 · Adversarial robustness and security › Certified and empirical robustness · approach: Robust overfitting and regularization · 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
Use Absum instead of standard weight decayCNNs · MNIST, FMNIST, and SVHN under SFAaverage robust accuracy: 98.59 -> 98.64; 83.09 -> 83.11; 93.20 -> 93.34Improved
Use Absum instead of L1 regularizationCNNs · MNIST, FMNIST, CIFAR10, CIFAR100, and SVHN under SFAaverage robust accuracy: 98.48 -> 98.64; 82.49 -> 83.11; 66.44 -> 79.05; 38.99 -> 48.69; 91.14 -> 93.34Mixed
Use spectral norm constraint instead of AbsumCNNs · CIFAR10 and CIFAR100 under SFAaverage robust accuracy: 79.05 -> 85.57; 48.69 -> 60.42Improved
Use Absum instead of spectral norm constraintCNNs · MNIST, FMNIST, and SVHN under SFAaverage robust accuracy: 98.55 -> 98.64; 82.60 -> 83.11; 93.20 -> 93.34Improved
Use Absum instead of weight decayCNNs · MNIST, FMNIST, and SVHN under transferred PGDrobust accuracy: 48.94 -> 76.34; 3.46 -> 30.08; 40.49 -> 49.11Improved
Use Absum instead of L1 regularizationCNNs · SVHN under transferred PGD (ε=4/255)robust accuracy: 52.79 -> 49.11Hurt
Use Absum instead of weight decayCNNs · CIFAR10 and CIFAR100 under transferred PGD (ε=4/255)robust accuracy: 18.48 -> 26.29; 17.40 -> 18.57Improved
Use Absum instead of L1 regularizationCNNs · SVHN under High-Frequency attackrobust accuracy: 17.03 -> 52.95Improved
Use Absum instead of weight decayCNNs · CIFAR10 and CIFAR100 under High-Frequency attackrobust accuracy: 52.82 -> 64.51; 36.15 -> 41.44Improved
Add Absum regularization to adversarial trainingCNNs trained with PGD adversarial training · MNIST and FMNISTrobust accuracyImproved
Add Absum regularization to adversarial trainingCNNs trained with PGD adversarial training · CIFAR10 and CIFAR100robust accuracyMixed
Add Absum regularization to naive trainingCNNs without adversarial training · PGD attacksrobust accuracyNo clear effect

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