AI Paper Atlas

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

Shaping Deep Feature Space Towards Gaussian Mixture for Visual Classification

Weitao Wan, Yu Cheng, Jiansheng Chen, Tong Wu, Yuanyi Zhong, Ming–Hsuan Yang

TPAMI 2022 · Adversarial robustness and security › Certified and empirical robustness · approach: Robust overfitting and regularization · 17 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 softmax loss with GM lossResNet-110 · CIFAR-10error rate (%): 6.43 -> 4.96Improved
replace softmax loss with GM lossResNet-50 · ImageNet val top-1error rate (%): 24.7 -> 23.28Improved
replace softmax loss with GM loss13-layer CNN · CIFAR-100 (no augmentation, C100)error rate (%): 25.61 -> 23.04Improved
add classification margin (α>0) to GM loss (ablation)GM loss · MNIST (100-D features)error rate (%): 0.49 -> 0.39Improved
add classification margin (α>0) to GM loss (ablation)GM loss · LFW face verificationaccuracy (%): 98.80 -> 99.41Improved
add likelihood regularization (λ>0) (ablation)GM loss · LFW face verificationaccuracy (%): 98.88 -> 99.41Improved
increase margin α on data without augmentation (ablation)GM loss · CIFAR-100 (no augmentation, C100)error rate (%): 23.04 -> 23.16Mixed
increase margin α with augmentation (ablation)GM loss · CIFAR-100 (augmentation, C100+)error rate (%): 20.94 -> 20.76Improved
increase margin α from 0 to 0.3 with augmentation layers (ablation)GM loss · CIFAR-100 with augmentation layers (λ=0.1)accuracy (%): 78.62 -> 79.24Improved
add GM loss to K-density detectorResNet-32 · CIFAR-10 white-box C&W-wb attackdetection AUC-ROC (%): 53.8 -> 84.3Improved
add GM loss to K-density detectorResNet · CIFAR-10 black-box C&W-wb attackdetection AUC-ROC (%): 75.0 -> 94.7Improved
add GM loss to K-density detectorResNet-32 · MNIST semi white-box BIMdetection AUC-ROC (%): 90.3 -> 99.7Improved

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