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

arXiv: https://arxiv.org/abs/2006.08545 · HTML: https://mlsanity.com/paper/2006.08545/

## Key facts

- Cited by 221 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 15 Jun 2020.

## Built on by (most cited first, 30 of 221)

- [Generalized Out-of-Distribution Detection: A Survey](https://mlsanity.com/paper/2110.11334/) (2021)
- [CutPaste: Self-Supervised Learning for Anomaly Detection and Localization](https://mlsanity.com/paper/2104.04015/) (2021)
- [ReAct: Out-of-distribution Detection With Rectified Activations](https://mlsanity.com/paper/2111.12797/) (2021)
- [On the Importance of Gradients for Detecting Distributional Shifts in the Wild](https://mlsanity.com/paper/2110.00218/) (2021)
- [CFLOW-AD: Real-Time Unsupervised Anomaly Detection with Localization via Conditional Normalizing Flows](https://mlsanity.com/paper/2107.12571/) (2021)
- [Offline Reinforcement Learning with Fisher Divergence Critic Regularization](https://mlsanity.com/paper/2103.08050/) (2021)
- [Delving into Out-of-Distribution Detection with Vision-Language Representations](https://mlsanity.com/paper/2211.13445/) (2022)
- [Low-Light Image Enhancement with Normalizing Flow](https://mlsanity.com/paper/2109.05923/) (2021)
- [OpenOOD v1.5: Enhanced Benchmark for Out-of-Distribution Detection](https://mlsanity.com/paper/2306.09301/) (2023)
- [How to Exploit Hyperspherical Embeddings for Out-of-Distribution Detection?](https://mlsanity.com/paper/2203.04450/) (2022)
- [Score-Based Diffusion Models as Principled Priors for Inverse Imaging](https://mlsanity.com/paper/2304.11751/) (2023)
- [Fully Convolutional Cross-Scale-Flows for Image-based Defect Detection](https://mlsanity.com/paper/2110.02855/) (2021)
- [LoCoOp: Few-Shot Out-of-Distribution Detection via Prompt Learning](https://mlsanity.com/paper/2306.01293/) (2023)
- [DEUP: Direct Epistemic Uncertainty Prediction](https://mlsanity.com/paper/2102.08501/) (2021)
- [Understanding Failures in Out-of-Distribution Detection with Deep Generative Models](https://mlsanity.com/paper/2107.06908/) (2021)
- [Types of Out-of-Distribution Texts and How to Detect Them](https://mlsanity.com/paper/2109.06827/) (2021)
- [Graph Posterior Network: Bayesian Predictive Uncertainty for Node Classification](https://mlsanity.com/paper/2110.14012/) (2021)
- [PyramidFlow: High-Resolution Defect Contrastive Localization using Pyramid Normalizing Flow](https://mlsanity.com/paper/2303.02595/) (2023)
- [Unknown-Aware Object Detection: Learning What You Don't Know from Videos in the Wild](https://mlsanity.com/paper/2203.03800/) (2022)
- [On the Impact of Spurious Correlation for Out-of-distribution Detection](https://mlsanity.com/paper/2109.05642/) (2021)
- [Normalizing Flows for Human Pose Anomaly Detection](https://mlsanity.com/paper/2211.10946/) (2022)
- [Unsupervised Out-of-Distribution Detection with Diffusion Inpainting](https://mlsanity.com/paper/2302.10326/) (2023)
- [RankFeat: Rank-1 Feature Removal for Out-of-distribution Detection](https://mlsanity.com/paper/2209.08590/) (2022)
- [Boosting Out-of-distribution Detection with Typical Features](https://mlsanity.com/paper/2210.04200/) (2022)
- [A Closer Look at the Robustness of Contrastive Language-Image Pre-Training (CLIP)](https://mlsanity.com/paper/2402.07410/) (2024)
- [On Advantages of Mask-level Recognition for Outlier-aware Segmentation](https://mlsanity.com/paper/2301.03407/) (2023)
- [A Geometric View of Data Complexity: Efficient Local Intrinsic Dimension Estimation with Diffusion Models](https://mlsanity.com/paper/2406.03537/) (2024)
- [GL-MCM: Global and Local Maximum Concept Matching for Zero-Shot Out-of-Distribution Detection](https://mlsanity.com/paper/2304.04521/) (2023)
- [LINe: Out-of-Distribution Detection by Leveraging Important Neurons](https://mlsanity.com/paper/2303.13995/) (2023)
- [Outlier detection by ensembling uncertainty with negative objectness](https://mlsanity.com/paper/2402.15374/) (2024)

## Abstract (arXiv)

Detecting out-of-distribution (OOD) data is crucial for robust machine learning systems. Normalizing flows are flexible deep generative models that often surprisingly fail to distinguish between in- and out-of-distribution data: a flow trained on pictures of clothing assigns higher likelihood to handwritten digits. We investigate why normalizing flows perform poorly for OOD detection. We demonstrate that flows learn local pixel correlations and generic image-to-latent-space transformations which are not specific to the target image dataset. We show that by modifying the architecture of flow coupling layers we can bias the flow towards learning the semantic structure of the target data, improving OOD detection. Our investigation reveals that properties that enable flows to generate high-fidelity images can have a detrimental effect on OOD detection.