mlsanityData to 7 Oct 2026

How Does Unlabeled Data Provably Help Out-of-Distribution Detection?

arXiv 2402.03502 ↗First posted 5 Feb 2024cs.LG stat.MLcited by 36

Not on the map. Why?

Abstract (arXiv)

Using unlabeled data to regularize the machine learning models has demonstrated promise for improving safety and reliability in detecting out-of-distribution (OOD) data. Harnessing the power of unlabeled in-the-wild data is non-trivial due to the heterogeneity of both in-distribution (ID) and OOD data. This lack of a clean set of OOD samples poses significant challenges in learning an optimal OOD classifier. Currently, there is a lack of research on formally understanding how unlabeled data helps OOD detection. This paper bridges the gap by introducing a new learning framework SAL (Separate And Learn) that offers both strong theoretical guarantees and empirical effectiveness. The framework separates candidate outliers from the unlabeled data and then trains an OOD classifier using the candidate outliers and the labeled ID data. Theoretically, we provide rigorous error bounds from the lens of separability and learnability, formally justifying the two components in our algorithm. Our theory shows that SAL can separate the candidate outliers with small error rates, which leads to a generalization guarantee for the learned OOD classifier. Empirically, SAL achieves state-of-the-art performance on common benchmarks, reinforcing our theoretical insights. Code is publicly available at https://github.com/deeplearning-wisc/sal.

Key facts. Cited by 36 papers in the map (data through 7 Oct 2026). Builds on 78 of its 104 references that are in the map. First posted on arXiv 5 Feb 2024.

Builds on

78 of 104 references.

References · 78

  1. 2015Dropout as a Bayesian Approximation: Representing Model Uncertainty in Deep Learningcited by 4,416
  2. 2016Simple and Scalable Predictive Uncertainty Estimation using Deep Ensemblescited by 3,888
  3. 2016Wide Residual Networkscited by 3,057
  4. 2016A Baseline for Detecting Misclassified and Out-of-Distribution Examples in Neural Networkscited by 2,362
  5. 2018A Simple Unified Framework for Detecting Out-of-Distribution Samples and Adversarial Attackscited by 1,472
  6. 2015LSUN: Construction of a Large-scale Image Dataset using Deep Learning with Humans in the Loopcited by 1,350
  7. 2017Enhancing The Reliability of Out-of-distribution Image Detection in Neural Networkscited by 1,325
  8. 2020Energy-based Out-of-distribution Detectioncited by 1,296
  9. 2018Deep Anomaly Detection with Outlier Exposurecited by 939
  10. 2021Generalized Out-of-Distribution Detection: A Surveycited by 796
  11. 2014Deep Neural Networks are Easily Fooled: High Confidence Predictions for Unrecognizable Imagescited by 774
  12. 2022Out-of-Distribution Detection with Deep Nearest Neighborscited by 596
38 more
  1. 2018Predictive Uncertainty Estimation via Prior Networkscited by 555
  2. 2019A Simple Baseline for Bayesian Uncertainty in Deep Learningcited by 555
  3. 2019Benign Overfitting in Linear Regressioncited by 544
  4. 2015Towards Open Set Deep Networkscited by 533
  5. 2021ReAct: Out-of-distribution Detection With Rectified Activationscited by 486
  6. 2020CSI: Novelty Detection via Contrastive Learning on Distributionally Shifted Instancescited by 441
  7. 2017Training Confidence-calibrated Classifiers for Detecting Out-of-Distribution Samplescited by 389
  8. 2020Simple and Principled Uncertainty Estimation with Deterministic Deep Learning via Distance Awarenesscited by 389
  9. 2020BatchEnsemble: An Alternative Approach to Efficient Ensemble and Lifelong Learningcited by 368
  10. 2021On the Importance of Gradients for Detecting Distributional Shifts in the Wildcited by 338
  11. 2019SelectiveNet: A Deep Neural Network with an Integrated Reject Optioncited by 333
  12. 2018Why ReLU networks yield high-confidence predictions far away from the training data and how to mitigate the problemcited by 323
  13. 2022VOS: Learning What You Don't Know by Virtual Outlier Synthesiscited by 300
  14. 2021Exploring the Limits of Out-of-Distribution Detectioncited by 283
  15. 2022OpenOOD: Benchmarking Generalized Out-of-Distribution Detectioncited by 267
  16. 2021SSD: A Unified Framework for Self-Supervised Outlier Detectioncited by 264
  17. 2022Mitigating Neural Network Overconfidence with Logit Normalizationcited by 251
  18. 2020Being Bayesian, Even Just a Bit, Fixes Overconfidence in ReLU Networkscited by 227
  19. 2022Delving into Out-of-Distribution Detection with Vision-Language Representationscited by 215
  20. 2022Extremely Simple Activation Shaping for Out-of-Distribution Detectioncited by 212
  21. 2016Robust Estimators in High Dimensions without the Computational Intractabilitycited by 209
  22. 2021A Simple Fix to Mahalanobis Distance for Improving Near-OOD Detectioncited by 206
  23. 2022Out-of-Distribution Detection and Selective Generation for Conditional Language Modelscited by 193
  24. 2023OpenOOD v1.5: Enhanced Benchmark for Out-of-Distribution Detectioncited by 178
  25. 2021DICE: Leveraging Sparsification for Out-of-Distribution Detectioncited by 166
  26. 2023Non-Parametric Outlier Synthesiscited by 129
  27. 2022How to Exploit Hyperspherical Embeddings for Out-of-Distribution Detection?cited by 128
  28. 2018Sever: A Robust Meta-Algorithm for Stochastic Optimizationcited by 123
  29. 2022Is Out-of-Distribution Detection Learnable?cited by 109
  30. 2022POEM: Out-of-Distribution Detection with Posterior Samplingcited by 108
  31. 2021Semantically Coherent Out-of-Distribution Detectioncited by 107
  32. 2019Reverse KL-Divergence Training of Prior Networks: Improved Uncertainty and Adversarial Robustnesscited by 104
  33. 2019Recent Advances in Algorithmic High-Dimensional Robust Statisticscited by 100
  34. 2021Can multi-label classification networks know what they don't know?cited by 100
  35. 2022Training OOD Detectors in their Natural Habitatscited by 86
  36. 2023Dream the Impossible: Outlier Imagination with Diffusion Modelscited by 77
  37. 2022Benign Overfitting without Linearity: Neural Network Classifiers Trained by Gradient Descent for Noisy Linear Datacited by 74
  38. 2022Unknown-Aware Object Detection: Learning What You Don't Know from Videos in the Wildcited by 70

50 most cited of 78 All 78 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 16; 2025: 15; 2026: 5.

36 papers.

Sort:

By area:

  1. 2024HaloScope: Harnessing Unlabeled LLM Generations for Hallucination Detectioncited by 96
  2. 2024Process Reward Model with Q-Value Rankingscited by 78
  3. 2024Out-of-Distribution Detection: A Task-Oriented Survey of Recent Advancescited by 24
  4. 2025GOLD: Graph Out-of-Distribution Detection via Implicit Adversarial Latent Generationcited by 15
  5. 2024Gradient-Regularized Out-of-Distribution Detectioncited by 13
  6. 2024Can OOD Object Detectors Learn from Foundation Models?cited by 10
  7. 2024On the Learnability of Out-of-distribution Detectioncited by 8
  8. 2025Noisy Test-Time Adaptation in Vision-Language Modelscited by 8
  9. 2024AHA: Human-Assisted Out-of-Distribution Generalization and Detectioncited by 7
  10. 2025AdaSCALE: Adaptive Scaling for OOD Detectioncited by 7
  11. 2024When and How Does In-Distribution Label Help Out-of-Distribution Detection?cited by 6
  12. 2024Out-of-Distribution Learning with Human Feedbackcited by 6
24 more
  1. 2024Adversarially Robust Out-of-Distribution Detection Using Lyapunov-Stabilized Embeddingscited by 6
  2. 2025Efficient Multivariate Robust Mean Estimation Under Mean-Shift Contaminationcited by 5
  3. 2025Graph Synthetic Out-of-Distribution Exposure with Large Language Modelscited by 5
  4. 2025GOOD: Training-Free Guided Diffusion Sampling for Out-of-Distribution Detectioncited by 5
  5. 2024What If the Input is Expanded in OOD Detection?cited by 4
  6. 2025Out-of-Distribution Detection using Synthetic Data Generationcited by 4
  7. 2025BOOD: Boundary-based Out-Of-Distribution Data Generationcited by 4
  8. 2024Bridging OOD Detection and Generalization: A Graph-Theoretic Viewcited by 3
  9. 2025Can We Ignore Labels In Out of Distribution Detection?cited by 3
  10. 2024FEVER-OOD: Free Energy Vulnerability Elimination for Robust Out-of-Distribution Detectioncited by 2
  11. 2025Open-Insect: Benchmarking Open-Set Recognition of Novel Species in Biodiversity Monitoringcited by 2
  12. 2025Medix: Out-of-Distribution Detection from Unlabeled Wild Data via Robust Gradient Statisticscited by 2
  13. 2025Harnessing Feature Resonance under Arbitrary Target Alignment for Out-of-Distribution Node Detectioncited by 1
  14. 2025LoD: Loss-difference OOD Detection by Intentionally Label-Noisifying Unlabeled Wild Datacited by 1
  15. 2025Gradient Short-Circuit: Efficient Out-of-Distribution Detection via Feature Interventioncited by 1
  16. 2025Confidence and Dispersity as Signals: Unsupervised Model Evaluation and Rankingcited by 1
  17. 2026Geometrically Constrained Outlier Synthesiscited by 1
  18. 2026What Information Matters? Graph Out-of-Distribution Detection via Tri-Component Information Decompositioncited by 1
  19. 2024Iterative Deployment Exposure for Unsupervised Out-of-Distribution Detectioncited by 0
  20. 2024Resultant: Incremental Effectiveness on Likelihood for Unsupervised Out-of-Distribution Detectioncited by 0
  21. 2024Semantic or Covariate? A Study on the Intractable Case of Out-of-Distribution Detectioncited by 0
  22. 2026The Invisible Gorilla Effect in Out-of-distribution Detectioncited by 0
  23. 2026Sample Complexity Bounds for Robust Mean Estimation with Mean-Shift Contaminationcited by 0
  24. 2026Self-Poisoning in Adaptive Out-of-Distribution Detection: A Sharp-Threshold Theory and Certified Label-Free Calibrationcited by 0