AI Paper Atlas

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

Mean-Shifted Contrastive Loss for Anomaly Detection

Tal Reiss, Yedid Hoshen

AAAI 2023 · Label-efficient and robust learning › OOD and anomaly detection · approach: Unified multi-class anomaly detection · 88 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-pretrained features with mean-shifted contrastive loss instead of PANDA center lossResNet152 anomaly detection · CIFAR-10ROC-AUC: 96.2 -> 97.2Improved
Replace standard contrastive loss with mean-shifted contrastive lossImageNet-pretrained feature adaptation · One-class anomaly detectionROC-AUCImproved
Add angular center loss to MSC (ablation)ResNet152 anomaly detection · CIFAR-10ROC-AUC: 97.2 -> 97.5Improved
Replace Euclidean center loss with angular center loss (ablation)PANDA · CIFAR-10ROC-AUC: 96.2 -> 96.8Improved
Normalize pretrained features to remove confidence magnitude (ablation)ImageNet-pretrained feature anomaly detection · CIFAR-10ROC-AUC: 92.5 -> 95.8Improved
Use mean-shifted contrastive loss across different pretrained backbonesImageNet-pretrained anomaly detection · CIFAR-10ROC-AUC: 89.3 -> 97.0; 85.6 -> 95.7; 95.7 -> 98.6Improved
Initialize MSC without a strong pretrained feature prior (ablation)MSC feature adaptation · Anomaly detectionROC-AUCHurt
Use ImageNet-pretrained features with CSI rather than the standard CSI setup (ablation)CSI · CIFAR-10ROC-AUC: 94.3 -> 89.5Hurt
Increase the CSI backbone size from ResNet18 to larger backbones (ablation)CSI · CIFAR-10ROC-AUC: 94.3%No clear effect
Compress the training gallery with k-means instead of using the full training set (ablation)MSC anomaly scoring · CIFAR-10 OCCROC-AUC: 97.2 -> 94.2; 97.2 -> 95.8; 97.2 -> 96.1; 97.2 -> 97.0Mixed
Train standard contrastive loss with smaller temperature (ablation)ImageNet-pretrained contrastive adaptation · CIFAR-10ROC-AUCHurt
Use the mean-shifted objective on multi-class normal dataPretrained feature adaptation · CIFAR-10 multi-class ADROC-AUC: 79.0 -> 85.3Improved

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