AI Paper Atlas

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

Exploring the Limits of Out-of-Distribution Detection

Stanislav Fort, Jie Ren, Balaji Lakshminarayanan

NeurIPS 2021 · Label-efficient and robust learning › OOD and anomaly detection · approach: Spurious-aware OOD detection · 103 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
fine-tune ImageNet-21k pre-trained ViT instead of prior SOTAnear-OOD detection · CIFAR-100 vs CIFAR-10AUROC: 85% -> 96%Improved
transformer (ViT) instead of ResNet (BiT) backbone with same pre-training data (ablation)fine-tuned OOD detector · CIFAR-100 vs CIFAR-10AUROC (Mahalanobis): 90.10 -> 95.53Improved
Mahalanobis score instead of MSP (ablation)fine-tuned ViT-B_16 · CIFAR-100 vs CIFAR-10AUROC: 91.89 -> 95.53Improved
supervised instead of self-supervised (DINO) pre-training at matched accuracy (ablation)ViT-B_16 · CIFAR-100 vs CIFAR-10AUROC (Mahalanobis): 88.78 -> 93.05Improved
scale up to ViT-L_16 (ablation)ViT · CIFAR-100 vs CIFAR-10AUROC (Mahalanobis): 96.23 -> 97.98Improved
add few-shot outlier exposure (1 example per OOD class)fine-tuned R50+ViT · CIFAR-100 vs CIFAR-10AUROC: 96.23 -> 98.70Improved
fine-tune on in-distribution before outlier exposure (ablation)R50+ViT, 1-shot OE · CIFAR-100 vs CIFAR-10AUROC: 88.73 -> 98.70Improved
unsupervised BERT pre-training + fine-tuning instead of 1D CNNgenomics classifier · Genomics OODAUROC (Mahalanobis): 64.75 -> 77.49Improved
add few-shot outlier exposureBERT genomics · Genomics OODAUROC: 76.73 -> 88.48Improved
use outlier class labels instead of collapsing to single OOD class (ablation)fine-tuned BERT OE · Genomics OODAUROC: 84.36 -> 82.21 (1-shot); 87.94 -> 88.49 (100-shot)Mixed
add OOD class names as candidate labelsCLIP zero-shot · CIFAR-100 vs CIFAR-10AUROC: 69.49 -> 94.68Improved
use 'the opposite of' + in-dist names as OOD labels (ablation)CLIP zero-shot · CIFAR-100 vs CIFAR-10AUROC: 69.49 -> 61.37Hurt

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