AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

Learn to Rank: Visual Attribution by Learning Importance Ranking

David Schinagl, Christian Fruhwirth-Reisinger, Alexander Prutsch, Samuel Schulter, Horst Possegger

ECCV 2026 · Architectures and efficiency › RNNs and normalization · approach: Differentiable optimization layers · 0 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
add 3 test-time refinement steps (ablation)AHA · ImageNet val, ViT-B/16 classifier, predicted classDeletion AUC: 0.1381 -> 0.1215Improved
add 3 test-time refinement steps (ablation)AHA · ImageNet val, ViT-B/16 classifier, predicted classInsertion AUC: 0.6112 -> 0.6485Improved
increase training perturbation steps S from 4 to 16 (ablation)AHA · ImageNet val, ViT-B/16Insertion AUCImproved
increase training perturbation steps beyond 16 (ablation)AHA · ImageNet val, ViT-B/16Insertion AUCNo clear effect
sample grid size G from U(7,28) instead of fixed G=7 (ablation)AHA · ViT-B/16 hyperparameter ablationRISE Ins.-Del.: 0.671 -> 0.715Improved
vary sorting temperature tau (ablation)AHA · ViT-B/16 hyperparameter ablationRISE Ins.-Del.: 0.703 to 0.715No clear effect
optimize deletion/insertion directly with learned amortized explainer instead of per-sample optimization (MDA)AHA · MDA protocol (5,000 images), ViT-B/16Insertion - Deletion: 0.624 -> 0.716Improved
use learned explainer instead of TIS on insertion at T=0AHA · ImageNet val, ViT-B/16 classifier, predicted classInsertion AUC: 0.6381 -> 0.6112Hurt
apply explainer trained on ViT-B/32 to ViT-B/16 with T=3 refinement (ablation)AHA · Cross-architecture transfer B/32 to B/16 (RISE Ins.-Del.)RISE Ins.-Del.: 0.650 -> 0.739Improved
apply AHA to ConvNeXt without architecture changesAHA · ImageNet val, ConvNeXt Small classifier, predicted classDeletion AUC: 0.2748 -> 0.1741Improved
AHA T=0 versus MDA on segmentation localizationAHA · ImageNet segmentation, ViT-B/32IoU: 0.702 -> 0.667Hurt

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