Input complexity and out-of-distribution detection with likelihood-based generative models

Benchmark Model Rank Results
anomaly-detection-on-unlabeled-cifar-10-vs-cifar-100Input Complexity (Glow)#5AUROC: 73.6
anomaly-detection-on-unlabeled-cifar-10-vs-cifar-100Likelihood (Glow)#6AUROC: 58.2
anomaly-detection-on-unlabeled-cifar-10-vs-cifar-100Input Complexity (PixelCNN++)#7AUROC: 53.5
anomaly-detection-on-unlabeled-cifar-10-vs-cifar-100Likelihood (PixelCNN++)#8AUROC: 52.6