AI Paper Atlas

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

Reliable Evaluation of Attribution Maps in CNNs: A Perturbation-Based Approach

Lars Nieradzik, Henrike Stephani, Janis Keuper

IJCV 2024 · Adversarial robustness and security › Certified and empirical robustness · approach: Robust overfitting and regularization · 12 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 pixel masking with FGSM adversarial perturbation in insertion/deletionattribution map evaluation · Uniform last-place sanity check (15 cases)cases passed: 11 -> 15Improved
replace pixel masking with FGSM adversarial perturbationattribution map evaluation · Canny second-to-last (15 cases)cases passed: 3 -> 15Improved
replace deletion with perturbation metricattribution map evaluation · ranking consistency across 15 combosKendall's tau: 0.236 -> 0.466Improved
replace insertion with perturbation metricattribution map evaluation · ranking consistency across 15 combosKendall's tau: 0.353 -> 0.466Improved
replace blurred insertion with perturbation metricattribution map evaluation · ranking consistencyKendall's tau: 0.432 -> 0.466Improved
replace insertion/deletion with perturbation metricattribution map evaluation · monotonicitymonotonicity: 0.671 -> 0.967Improved
replace insertion/deletion with perturbation metricattribution map evaluation · smoothness (lower better)smoothness: 1.658 -> 0.891Improved
replace blurred insertion with perturbation metricattribution map evaluation · smoothness (lower better)smoothness: 1.238 -> 0.891Improved
use stronger PGD attack instead of FGSM (ablation)perturbation metric · ChestX-ray8attack success rate: 95.80% -> 100%Improved
use stronger PGD attack instead of FGSM (ablation)perturbation metric · ranking of attribution mapsrankingNo clear effect
increase epsilon beyond 1/255 (ablation)perturbation metric · monotonicitymonotonicityHurt

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