AI Paper Atlas

Research map / Probabilistic and scientific modeling / Variational and generative sampling

Why Normalizing Flows Fail to Detect Out-of-Distribution Data

Polina Kirichenko, Pavel Izmailov, Andrew G Wilson

NeurIPS 2020 · Probabilistic and scientific modeling › Variational and generative sampling · approach: Normalizing flow training and design · 30 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
Train RealNVP and Glow on raw image pixels using maximum likelihoodRealNVP and Glow · CIFAR-10, CelebA, and SVHN OOD pairsAUROC: 56.0; 6.0; 98.4; 67.7; 6.3; 99.0Mixed
Train the flow on EfficientNet semantic image embeddings rather than raw pixelsRealNVP · CelebA, CIFAR-10, and SVHN embeddingsAUROC: 99.99; 99.99; 99.99; 73.31; 100.0; 99.98Improved
Use a cycle-mask instead of checkerboard or horizontal masks (ablation)RealNVP · FashionMNIST (OOD: MNIST)relative likelihood rankingImproved
Use a cycle-mask instead of checkerboard or horizontal masks (ablation)RealNVP · CelebA (OOD: SVHN)relative likelihood rankingMixed
Add a low-dimensional bottleneck to the st-networks (ablation)RealNVP · FashionMNIST (OOD: MNIST)relative likelihood rankingImproved
Add a low-dimensional bottleneck to the st-networks (ablation)RealNVP · CelebA (OOD: SVHN)relative likelihood rankingMixed
Train with a maximum-likelihood objective that also minimizes likelihood on a specified OOD datasetRealNVP · CIFAR-10 vs CelebA, CIFAR-10 vs SVHN, and FashionMNIST vs MNISTlikelihood separationImproved
Minimize likelihood on one selected OOD datasetRealNVP · Other unseen OOD datasetslikelihood separationNo clear effect

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