AI Paper Atlas

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

Learning Non-Linear Invariants for Unsupervised Out-of-Distribution Detection

Lars Doorenbos, Raphael Sznitman, Pablo Márquez Neila

ECCV 2024 · Label-efficient and robust learning › OOD and anomaly detection · approach: Spurious-aware OOD detection · 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
replace linear (PCA) invariants with non-linear volume-preserving network invariants (ablation)MahaAD · General U-OOD benchmarkmean AUC: 86.5 -> 87.2Improved
add backward (reconstruction) loss (ablation)NL-Invs · General U-OOD benchmarkmean AUC: 86.9 -> 87.2Improved
add 2-NN score to invariant score (ablation)NL-Invs · General U-OOD benchmarkmean AUC: 87.2 -> 87.9Improved
use non-linear instead of affine invariants on tabular dataNL-Invs vs MahaAD · Shallow tabular U-OODmean AUC: 82.9 -> 93.5 (10.6 points)Improved
use non-linear invariants on deep image features vs affineNL-Invs vs MahaAD · General U-OOD benchmarkmean AUC: 86.5 -> 87.9Improved
vary hyperparameter p from 0.5 to 10 (ablation)NL-Invs · General U-OOD (ResNet18)AUC: 86.5 to 86.8No clear effect
use stronger pre-trained backbone (ConvNeXT-B vs ResNet18) (ablation)NL-Invs · uni-classAUC: 87.8 -> 94.4Improved
NL-Invs vs best baseline on uni-ano and uni-classNL-Invs · uni-ano / uni-classAUC: 85.8 vs 95.7 (CFlow); 93.3 vs 96.3 (MSCL)Hurt
pre-trained features for shift-low-res (CIFAR10:SVHN)NL-Invs · shift-low-resAUC: 97.8Improved

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