Semantic or Covariate? A Study on the Intractable Case of Out-of-Distribution Detection
Not on the map. Why?
Abstract (arXiv)
The primary goal of out-of-distribution (OOD) detection tasks is to identify inputs with semantic shifts, i.e., if samples from novel classes are absent in the in-distribution (ID) dataset used for training, we should reject these OOD samples rather than misclassifying them into existing ID classes. However, we find the current definition of "semantic shift" is ambiguous, which renders certain OOD testing protocols intractable for the post-hoc OOD detection methods based on a classifier trained on the ID dataset. In this paper, we offer a more precise definition of the Semantic Space and the Covariate Space for the ID distribution, allowing us to theoretically analyze which types of OOD distributions make the detection task intractable. To avoid the flaw in the existing OOD settings, we further define the "Tractable OOD" setting which ensures the distinguishability of OOD and ID distributions for the post-hoc OOD detection methods. Finally, we conduct several experiments to demonstrate the necessity of our definitions and validate the correctness of our theorems.
Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 29 of its 31 references that are in the map. First posted on arXiv 18 Nov 2024.
Builds on
29 of 31 references.
References · 29
- 2015Deep Residual Learning for Image Recognitioncited by 52,096
- 2016A Baseline for Detecting Misclassified and Out-of-Distribution Examples in Neural Networkscited by 2,385
- 2018A Simple Unified Framework for Detecting Out-of-Distribution Samples and Adversarial Attackscited by 1,485
- 2017Enhancing The Reliability of Out-of-distribution Image Detection in Neural Networkscited by 1,328
- 2020Energy-based Out-of-distribution Detectioncited by 1,299
- 2018Deep Anomaly Detection with Outlier Exposurecited by 942
- 2021Generalized Out-of-Distribution Detection: A Surveycited by 802
- 2022Out-of-Distribution Detection with Deep Nearest Neighborscited by 600
- 2020Generalized ODIN: Detecting Out-of-distribution Image without Learning from Out-of-distribution Datacited by 408
- 2021On the Importance of Gradients for Detecting Distributional Shifts in the Wildcited by 339
- 2022VOS: Learning What You Don't Know by Virtual Outlier Synthesiscited by 300
- 2022OpenOOD: Benchmarking Generalized Out-of-Distribution Detectioncited by 268
17 more
- 2022Extremely Simple Activation Shaping for Out-of-Distribution Detectioncited by 212
- 2021A Simple Fix to Mahalanobis Distance for Improving Near-OOD Detectioncited by 207
- 2021DICE: Leveraging Sparsification for Out-of-Distribution Detectioncited by 166
- 2022How to Exploit Hyperspherical Embeddings for Out-of-Distribution Detection?cited by 128
- 2019Unsupervised Out-of-Distribution Detection by Maximum Classifier Discrepancycited by 112
- 2022Is Out-of-Distribution Detection Learnable?cited by 109
- 2021Semantically Coherent Out-of-Distribution Detectioncited by 107
- 2021Understanding Failures in Out-of-Distribution Detection with Deep Generative Modelscited by 89
- 2021Provable Guarantees for Understanding Out-of-distribution Detectioncited by 78
- 2023Scaling for Training Time and Post-hoc Out-of-distribution Detection Enhancementcited by 77
- 2022Full-Spectrum Out-of-Distribution Detectioncited by 63
- 2022RankFeat: Rank-1 Feature Removal for Out-of-distribution Detectioncited by 52
- 2024Learning with Mixture of Prototypes for Out-of-Distribution Detectioncited by 43
- 2024How Does Unlabeled Data Provably Help Out-of-Distribution Detection?cited by 36
- 2023ImageNet-OOD: Deciphering Modern Out-of-Distribution Detection Algorithmscited by 22
- 2023Neural Relation Graph: A Unified Framework for Identifying Label Noise and Outlier Datacited by 20
- 2024When and How Does In-Distribution Label Help Out-of-Distribution Detection?cited by 6
Built on by
Not cited yet.