CSI: Novelty Detection via Contrastive Learning on Distributionally Shifted Instances

Benchmark Model Rank Results
anomaly-detection-on-unlabeled-imagenet-30-vs-flowers-102CSI#1ROC-AUC: 94.7Network: ResNet-18
anomaly-detection-on-unlabeled-cifar-10-vs-cifar-100CSI#3AUROC: 89.3Network: ResNet-18
anomaly-detection-on-unlabeled-cifar-10-vs-lsun-fixCSI#3ROC-AUC: 90.3Network: ResNet-18
anomaly-detection-on-unlabeled-imagenet-30-vs-cub-200CSI#3ROC-AUC: 71.5Network: ResNet-18
anomaly-detection-on-one-class-imagenet-30CSI#4AUROC: 91.6
anomaly-detection-on-one-class-cifar-100CSI#6AUROC: 89.6
anomaly-detection-on-one-class-cifar-10CSI#11AUROC: 94.3