InFlow: Robust outlier detection utilizing Normalizing Flows
Not on the map. Why?
Abstract (arXiv)
Normalizing flows are prominent deep generative models that provide tractable probability distributions and efficient density estimation. However, they are well known to fail while detecting Out-of-Distribution (OOD) inputs as they directly encode the local features of the input representations in their latent space. In this paper, we solve this overconfidence issue of normalizing flows by demonstrating that flows, if extended by an attention mechanism, can reliably detect outliers including adversarial attacks. Our approach does not require outlier data for training and we showcase the efficiency of our method for OOD detection by reporting state-of-the-art performance in diverse experimental settings. Code available at https://github.com/ComputationalRadiationPhysics/InFlow .
Key facts. Cited by 3 papers in the map (data through 7 Oct 2026). Builds on 32 of its 44 references that are in the map. First posted on arXiv 10 Jun 2021.
Builds on
32 of 44 references.
References · 32
- 2017Fashion-MNIST: a Novel Image Dataset for Benchmarking Machine Learning Algorithmscited by 4,283
- 2016Simple and Scalable Predictive Uncertainty Estimation using Deep Ensemblescited by 3,888
- 2014Deep Learning Face Attributes in the Wildcited by 3,679
- 2019Benchmarking Neural Network Robustness to Common Corruptions and Perturbationscited by 2,857
- 2016A Baseline for Detecting Misclassified and Out-of-Distribution Examples in Neural Networkscited by 2,362
- 2016Density estimation using Real NVPcited by 2,029
- 2015Deep Reinforcement Learning with Double Q-learningcited by 1,704
- 2018Glow: Generative Flow with Invertible 1x1 Convolutionscited by 1,539
- 2018A Simple Unified Framework for Detecting Out-of-Distribution Samples and Adversarial Attackscited by 1,472
- 2017Enhancing The Reliability of Out-of-distribution Image Detection in Neural Networkscited by 1,325
- 2014NICE: Non-linear Independent Components Estimationcited by 1,052
- 2018Deep Anomaly Detection with Outlier Exposurecited by 939
20 more
- 2014Deep Neural Networks are Easily Fooled: High Confidence Predictions for Unrecognizable Imagescited by 774
- 2015Dueling Network Architectures for Deep Reinforcement Learningcited by 769
- 2019Neural Spline Flowscited by 590
- 2018FFJORD: Free-form Continuous Dynamics for Scalable Reversible Generative Modelscited by 584
- 2017Countering Adversarial Images using Input Transformationscited by 517
- 2019Using Self-Supervised Learning Can Improve Model Robustness and Uncertaintycited by 511
- 2019Likelihood Ratios for Out-of-Distribution Detectioncited by 451
- 2018Do Deep Generative Models Know What They Don't Know?cited by 443
- 2018GANomaly: Semi-Supervised Anomaly Detection via Adversarial Trainingcited by 409
- 2017Training Confidence-calibrated Classifiers for Detecting Out-of-Distribution Samplescited by 389
- 2018Learning Confidence for Out-of-Distribution Detection in Neural Networkscited by 245
- 2018Failing Loudly: An Empirical Study of Methods for Detecting Dataset Shiftcited by 240
- 2018Analyzing Inverse Problems with Invertible Neural Networkscited by 239
- 2020Why Normalizing Flows Fail to Detect Out-of-Distribution Datacited by 220
- 2019Input complexity and out-of-distribution detection with likelihood-based generative modelscited by 186
- 2020Likelihood Regret: An Out-of-Distribution Detection Score For Variational Auto-encodercited by 145
- 2020Disentanglement by Nonlinear ICA with General Incompressible-flow Networks (GIN)cited by 91
- 2020Density of States Estimation for Out-of-Distribution Detectioncited by 74
- 2020Robust Out-of-distribution Detection for Neural Networkscited by 57
- 2020CD-UAP: Class Discriminative Universal Adversarial Perturbationcited by 27
Built on by
Citing papers per year: 2022: 1; 2023: 1; 2024: 0; 2025: 0; 2026: 1.
3 papers.