AI Paper Atlas

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

WDiscOOD: Out-of-Distribution Detection via Whitened Linear Discriminant Analysis

Yiye Chen, Yunzhi Lin, Ruinian Xu, Patricio A. Vela

ICCV 2023 · Label-efficient and robust learning › OOD and anomaly detection · approach: Spurious-aware OOD detection · 9 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
combine discriminative and residual subspace scores (WDiscOOD) versus best baselinesWDiscOOD · ImageNet-1k OOD, ResNet-50AUROC: 86.64 -> 87.23Improved
apply WDiscOOD to contrastive encoders instead of Mahalanobis/KNNSupCon and CLIP encoders · ImageNet-1k OODAUROC: SupCon 90.35 -> 90.89; CLIP 75.31 -> 75.74Improved
use whitened discriminative residual score instead of principal residual score (ablation)Residual-subspace OOD score · ImageNet-1k OOD, ResNet-50AUROC: 84.01 -> 86.56Improved
add data whitening before LDA (ablation)WDiscOOD · ImageNet-1k OOD, ResNet-50FPR95: 74.56 -> 49.86Improved
add data whitening before LDA (ablation)WDiscOOD · ImageNet-1k OOD, ViT-B/16FPR95: 32.21 -> 26.49Improved
use Mahalanobis instead of Euclidean distance after whitened LDA (ablation)WDiscOOD · ImageNet-1k OOD, ResNet-50FPR95: 49.86 -> 49.85No clear effect
add Mahalanobis distance after LDA without whitening (ablation)WDiscOOD without whitening · ImageNet-1k OOD, ResNet-50FPR95: 74.56 -> 53.65Improved
combine WD and WDR subspace scores instead of a single subspace (ablation)WDiscOOD · ImageNet-1k OODFPR95Improved
use extreme score scale factor alpha (very small or large) (ablation)WDiscOOD · ImageNet-1k OOD, ResNet-50AUROCHurt
increase number of discriminants N_D (ablation)WDiscOOD · ImageNet-1k OOD, ResNet-50AUROCNo clear effect
use 20% of training data (200K) for LDA statistics (ablation)WDiscOOD · ImageNet-1k OOD, ResNet-50AUROCNo clear effect
apply feature-space methods on projection embedding instead of penultimate features (ablation)Mahalanobis, KNN, WDiscOOD · SupCon and CLIPFPR95: WDisc SupCon 40.10 -> 53.10; CLIP 77.57 -> 92.25Hurt

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