AI Paper Atlas

Research map / Label-efficient and robust learning / OOD and anomaly detection

RankFeat&RankWeight: Rank-1 Feature/Weight Removal for Out-of-Distribution Detection

Yue Song, Wei Wang, Nicu Sebe

TPAMI 2024 · Label-efficient and robust learning › OOD and anomaly detection · approach: Provable and feature-based OOD detection · 30 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 RankFeat+RankWeight togetherRankFeat (Block 3+4) · ImageNet-1k, ResNetv2-101average FPR95: 36.80 -> 16.13Improved
remove rank-1 matrix instead of keeping only rank-1 (ablation)RankFeat · ImageNet-1k, ResNetv2-101average FPR95: 58.49 -> 39.69Improved
remove rank-1 matrix instead of rank-2 (ablation)RankFeat · ImageNet-1k, ResNetv2-101average FPR95: 57.92 -> 39.69Improved
remove rank-1 matrix instead of rank-3 (ablation)RankFeat · ImageNet-1k, ResNetv2-101average FPR95: 61.94 -> 39.69Improved
apply RankFeat to Block 3 or 4 instead of Block 1 (ablation)RankFeat · ImageNet-1k, ResNetv2-101average FPR95: 76.65 -> 45.80Improved
add RankWeight to GradNormGradNorm · ImageNet-1k, ResNetv2-101average FPR95: 54.70 -> 29.99Improved
add RankWeight to ReActReAct · ImageNet-1k, ResNetv2-101average FPR95: 57.66 -> 36.12Improved
add RankWeight to VRAVRA · ImageNet-1k, ResNetv2-101average FPR95: 30.85 -> 17.75Improved
add RankWeight to ASHASH · ImageNet-1k, ResNetv2-101average FPR95: 31.44 -> 15.46Improved
use power iteration (10 iterations) instead of SVD (ablation)RankFeat · ImageNet-1k, ResNetv2-101average FPR95: 39.69 -> 39.88No clear effect
use RankWeight alone on ViT-B/16ViT-B/16 baseline · ImageNet-1k, ViT-B/16average FPR95: 60.75 -> 72.69Hurt
add RankFeat+RankWeight on SpeciesReAct · Species, ResNetv2-101average FPR95: 67.02 -> 22.30Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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