mlsanityData to 7 Oct 2026

On the Learnability of Out-of-distribution Detection

arXiv 2404.04865 ↗First posted 7 Apr 2024cs.LG cs.CV stat.MLcited by 8

Not on the map. Why?

Abstract (arXiv)

Supervised learning aims to train a classifier under the assumption that training and test data are from the same distribution. To ease the above assumption, researchers have studied a more realistic setting: out-of-distribution (OOD) detection, where test data may come from classes that are unknown during training (i.e., OOD data). Due to the unavailability and diversity of OOD data, good generalization ability is crucial for effective OOD detection algorithms, and corresponding learning theory is still an open problem. To study the generalization of OOD detection, this paper investigates the probably approximately correct (PAC) learning theory of OOD detection that fits the commonly used evaluation metrics in the literature. First, we find a necessary condition for the learnability of OOD detection. Then, using this condition, we prove several impossibility theorems for the learnability of OOD detection under some scenarios. Although the impossibility theorems are frustrating, we find that some conditions of these impossibility theorems may not hold in some practical scenarios. Based on this observation, we next give several necessary and sufficient conditions to characterize the learnability of OOD detection in some practical scenarios. Lastly, we offer theoretical support for representative OOD detection works based on our OOD theory.

Key facts. Cited by 8 papers in the map (data through 7 Oct 2026). Builds on 54 of its 110 references that are in the map. First posted on arXiv 7 Apr 2024.

Builds on

54 of 110 references.

References · 54

  1. 2017Enhancing The Reliability of Out-of-distribution Image Detection in Neural Networkscited by 1,325
  2. 2020Energy-based Out-of-distribution Detectioncited by 1,296
  3. 2021Generalized Out-of-Distribution Detection: A Surveycited by 796
  4. 2018Exploring the Limits of Weakly Supervised Pretrainingcited by 604
  5. 2022Out-of-Distribution Detection with Deep Nearest Neighborscited by 596
  6. 2021ReAct: Out-of-distribution Detection With Rectified Activationscited by 486
  7. 2017Snapshot Ensembles: Train 1, get M for freecited by 349
  8. 2021On the Importance of Gradients for Detecting Distributional Shifts in the Wildcited by 338
  9. 2022VOS: Learning What You Don't Know by Virtual Outlier Synthesiscited by 300
  10. 2022Mitigating Neural Network Overconfidence with Logit Normalizationcited by 251
  11. 2021MOS: Towards Scaling Out-of-distribution Detection for Large Semantic Spacecited by 233
  12. 2015Convergent Learning: Do different neural networks learn the same representations?cited by 215
38 more
  1. 2022Delving into Out-of-Distribution Detection with Vision-Language Representationscited by 215
  2. 2023OpenOOD v1.5: Enhanced Benchmark for Out-of-Distribution Detectioncited by 178
  3. 2021A Unified Survey on Anomaly, Novelty, Open-Set, and Out-of-Distribution Detection: Solutions and Future Challengescited by 154
  4. 2023Non-Parametric Outlier Synthesiscited by 129
  5. 2022How to Exploit Hyperspherical Embeddings for Out-of-Distribution Detection?cited by 128
  6. 2022Is Out-of-Distribution Detection Learnable?cited by 109
  7. 2022POEM: Out-of-Distribution Detection with Posterior Samplingcited by 108
  8. 2022Training OOD Detectors in their Natural Habitatscited by 86
  9. 2020Model Patching: Closing the Subgroup Performance Gap with Data Augmentationcited by 82
  10. 2021Provable Guarantees for Understanding Out-of-distribution Detectioncited by 78
  11. 2023Dream the Impossible: Outlier Imagination with Diffusion Modelscited by 77
  12. 2022Unknown-Aware Object Detection: Learning What You Don't Know from Videos in the Wildcited by 70
  13. 2021On the Impact of Spurious Correlation for Out-of-distribution Detectioncited by 68
  14. 2020Robust Out-of-distribution Detection for Neural Networkscited by 57
  15. 2023Feed Two Birds with One Scone: Exploiting Wild Data for Both Out-of-Distribution Generalization and Detectioncited by 57
  16. 2023Rethinking Domain Generalization for Face Anti-spoofing: Separability and Alignmentcited by 47
  17. 2020Frequency Domain Image Translation: More Photo-realistic, Better Identity-preservingcited by 46
  18. 2022Mitigating Memorization of Noisy Labels by Clipping the Model Predictioncited by 40
  19. 2023How Does Fine-Tuning Impact Out-of-Distribution Detection for Vision-Language Models?cited by 38
  20. 2023Learning to Augment Distributions for Out-of-Distribution Detectioncited by 38
  21. 2022Are Vision Transformers Robust to Spurious Correlations?cited by 36
  22. 2024How Does Unlabeled Data Provably Help Out-of-Distribution Detection?cited by 36
  23. 2023A Survey on Out-of-Distribution Detection in NLPcited by 31
  24. 2024Mitigating Fine-tuning based Jailbreak Attack with Backdoor Enhanced Safety Alignmentcited by 31
  25. 2022SoLar: Sinkhorn Label Refinery for Imbalanced Partial-Label Learningcited by 30
  26. 2023Is Fine-tuning Needed? Pre-trained Language Models Are Near Perfect for Out-of-Domain Detectioncited by 29
  27. 2018Understanding the Loss Surface of Neural Networks for Binary Classificationcited by 27
  28. 2022Out-of-distribution Detection via Frequency-regularized Generative Modelscited by 23
  29. 2023A Graph-Theoretic Framework for Understanding Open-World Semi-Supervised Learningcited by 22
  30. 2022PiCO+: Contrastive Label Disambiguation for Robust Partial Label Learningcited by 21
  31. 2023When and How Does Known Class Help Discover Unknown Ones? Provable Understanding Through Spectral Analysiscited by 21
  32. 2021Noise-robust Graph Learning by Estimating and Leveraging Pairwise Interactionscited by 20
  33. 2019Defense Against Adversarial Images using Web-Scale Nearest-Neighbor Searchcited by 19
  34. 2024Understanding the Learning Dynamics of Alignment with Human Feedbackcited by 15
  35. 2023Distributionally Robust Optimization with Probabilistic Groupcited by 14
  36. 2024Unsolvable Problem Detection: Robust Understanding Evaluation for Large Multimodal Modelscited by 12
  37. 2015Uncovering the Small Community Structure in Large Networks: A Local Spectral Approachcited by 10
  38. 2015Deep Manifold Traversal: Changing Labels with Convolutional Featurescited by 9

50 most cited of 54 All 54 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 3; 2025: 3; 2026: 2.

8 papers.

  1. 2024Process Reward Model with Q-Value Rankingscited by 78
  2. 2024Adaptive Hierarchical Graph Cut for Multi-granularity Out-of-distribution Detectioncited by 8
  3. 2025REALM: An MLLM-Agent Framework for Open World 3D Reasoning Segmentation and Editing on Gaussian Splattingcited by 6
  4. 2025LoD: Loss-difference OOD Detection by Intentionally Label-Noisifying Unlabeled Wild Datacited by 1
  5. 2024Double Descent Meets Out-of-Distribution Detection: Theoretical Insights and Empirical Analysis on the role of model complexitycited by 0
  6. 2025Predictive Sample Assignment for Semantically Coherent Out-of-Distribution Detectioncited by 0
  7. 2026On Hallucinations in Inverse Problems: Fundamental Limits and Provable Assessment Methodscited by 0
  8. 2026Surprises in Proper Positive-Only Learningcited by 0