AI Paper Atlas

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

UniCR: Universally Approximated Certified Robustness via Randomized Smoothing

Hanbin Hong, Binghui Wang, Yuan Hong

ECCV 2022 · Adversarial robustness and security › Certified and empirical robustness · approach: Randomized smoothing certification · 7 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
tune General Normal noise per input (I-OPT) instead of fixed LaplaceTeng's Laplace · CIFAR10 l1robustness score: 0.5606 -> 0.7027Improved
tune noise per input instead of fixed GaussianCohen's Gaussian · CIFAR10 l2robustness score: 0.5392 -> 0.7141Improved
tune noise per input instead of fixed GaussianYang's Gaussian · CIFAR10 linfrobustness score: 0.0098 -> 0.0136Improved
tune noise per input instead of fixed GaussianCohen's Gaussian · ImageNet l2robustness score: 1.0463 -> 1.0939Improved
tune noise per input instead of fixed GaussianYang's Gaussian · ImageNet linfrobustness score: 0.0027 -> 0.0031Improved
I-OPT on top 20% certified l1 radiusYang's Gaussian · ImageNet l1certified radius: 2.44 -> 2.36Hurt
I-OPT on top 100% l1 radiusYang's Gaussian · ImageNet l1certified radius: 0.95 -> 0.98Improved
I-OPT on linf radiusYang's Gaussian · ImageNet linfcertified radius x255: 1.60 -> 1.75Improved
lower beta from Laplace (1.00) to 0.75 in General Normal (ablation)Laplace noise · MNIST l1robustness score: 2.7448 -> 2.8354Improved
raise beta from Gaussian (2.00) to 2.25 (ablation)Gaussian noise · MNIST l2robustness score: 2.1983 -> 2.2081Improved
raise beta from 2.00 to 2.25 (ablation)Gaussian noise · MNIST linfrobustness score: 0.0879 -> 0.0880Improved
make beta very small (0.25) in General Normal (ablation)General Normal noise · MNIST l2robustness score: 2.2081 -> 0.0000Hurt

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