Pith. sign in

REVIEW 46 cited by

Deep Anomaly Detection with Outlier Exposure

Not yet reviewed by Pith; the record is open.

Re-run · record.json Download PDF Read on arXiv ↗

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 1812.04606 v3 pith:QJCZFLB4 submitted 2018-12-11 cs.LG cs.CLcs.CVstat.ML

Dan Hendrycks , Mantas Mazeika , Thomas Dietterich This is my paper · ORCID
classification cs.LGcs.CLcs.CVstat.ML
keywords anomalyexposureoutlierdeepdetectionanomalousauxiliarycifar-10
verification ladder T0 review T1 audit T2 compute T3 formal

Signed reviews

No signed human review yet.

0 comments
read the original abstract

It is important to detect anomalous inputs when deploying machine learning systems. The use of larger and more complex inputs in deep learning magnifies the difficulty of distinguishing between anomalous and in-distribution examples. At the same time, diverse image and text data are available in enormous quantities. We propose leveraging these data to improve deep anomaly detection by training anomaly detectors against an auxiliary dataset of outliers, an approach we call Outlier Exposure (OE). This enables anomaly detectors to generalize and detect unseen anomalies. In extensive experiments on natural language processing and small- and large-scale vision tasks, we find that Outlier Exposure significantly improves detection performance. We also observe that cutting-edge generative models trained on CIFAR-10 may assign higher likelihoods to SVHN images than to CIFAR-10 images; we use OE to mitigate this issue. We also analyze the flexibility and robustness of Outlier Exposure, and identify characteristics of the auxiliary dataset that improve performance.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 46 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. OpenAlex reports about 402 citations worldwide. Full citation record

  1. Teaching Models to Express Their Uncertainty in Words

    cs.CL 2022-05 unverdicted novelty 8.0 of 10

    GPT-3 can learn to express well-calibrated uncertainty about its answers using natural language phrases rather than logits.

  2. Beyond Binary Out-of-Distribution Detection: Characterizing Distributional Shifts with Multi-Statistic Diffusion Trajectories

    cs.LG 2025-10 unverdicted novelty 7.0 of 10

    DISC extracts multi-statistic trajectories from diffusion denoising to both detect and classify types of distributional shifts in OOD data.

  3. NOVA: A Benchmark for Anomaly Localization and Clinical Reasoning in Brain MRI

    eess.IV 2025-05 conditional novelty 7.0 of 10

    NOVA is a new evaluation-only benchmark of rare brain MRI pathologies where GPT-4o, Gemini 2.0 Flash, and Qwen2.5-VL-72B all exhibit large performance drops across anomaly localization, image captioning, and diagnosti...

  4. Likelihood Training of Cascaded Diffusion Models via Hierarchical Volume-preserving Maps

    cs.LG 2025-01 reject novelty 7.0 of 10

    A volume-preserving reparameterization makes the likelihood of cascaded diffusion models exactly computable, giving state-of-the-art density estimation on standard image benchmarks.

  5. ReliableNet: A Chance-Constrained Approach to Trustworthy Classification in Deep Learning

    cs.LG 2026-08 conditional novelty 6.0 of 10

    ReliableNet trains classifiers under an explicit budget on the joint probability of high-confidence and incorrect predictions, and reports held-out certification on six benchmarks.

  6. Representation Trajectories Matters: Complementary Evidence for OOD Detection and Image Classification

    cs.CV 2026-07 accept novelty 6.0 of 10

    Recording how an image's representation evolves block-by-block, relative to learned class routes, improves OOD detection in 131/152 comparisons and clean classification in 71/72 model–dataset cases.

  7. Learning from Noise: Effective-Rank Collapse and Out-of-Distribution Rejection in Restricted Boltzmann Machines

    cs.LG 2026-07 conditional novelty 6.0 of 10

    Auxiliary random-binary exposure collapses the effective rank of an RBM's J=WW^T toward the data covariance, enabling OOD rejection while preserving MNIST accuracy.

  8. Catching Disguised Transients with ASTRANet: Anomaly-Aware Spectroscopic Classification and Conformal Calibration

    astro-ph.IM 2026-07 conditional novelty 6.0 of 10

    ASTRANet combines a redshift-free spectral classifier, a 16-score anomaly detector, and conformal prediction to identify and calibrate uncertainty for out-of-taxonomy astronomical transients.

  9. Debiased Negative Mining Improves Out-of-distribution Detection with Pre-trained Vision-Language Models

    cs.LG 2026-05 unverdicted novelty 6.0 of 10

    Debiased negative mining via Monte-Carlo sampling from ID labels and unlabeled wild data improves OOD detection with VLMs and achieves new state-of-the-art results.

  10. GAMR: Geometric-Aware Manifold Regularization with Virtual Outlier Synthesis for Learning with Noisy Labels

    cs.CV 2026-05 unverdicted novelty 6.0 of 10

    GAMR introduces geometric-aware manifold regularization via virtual outlier synthesis to enhance intra-class compactness and inter-class separation, improving robustness to noisy labels beyond passive sample filtering.

  11. ROAST: Risk-aware Outlier-exposure for Adversarial Selective Training of Anomaly Detectors Against Evasion Attacks

    cs.CR 2026-03 unverdicted novelty 6.0 of 10

    ROAST selectively trains anomaly detectors on less vulnerable patient data with targeted outlier exposure, boosting recall by 16.2% in black-box settings and reducing training time by 88.3%.

  12. Force-Aware Residual DAgger via Trajectory Editing for Precision Insertion with Impedance Control

    cs.RO 2026-03 conditional novelty 6.0 of 10

    TER-DAgger uses force-prediction mismatches to trigger human corrections and residual-policy training, lifting precision-insertion success from 40.0% to 77.2% on average.

  13. Native Extrapolation Awareness in Flow-Based Conditional Generation

    cs.LG 2026-02 conditional novelty 6.0 of 10

    A contrastive flow-matching objective makes off-manifold conditions produce curved trajectories, so path curvature (the DOT score) separates invalid from valid inputs.

  14. LoFT: Parameter-Efficient Fine-Tuning for Long-tailed Semi-Supervised Learning in Open-World Scenarios

    cs.LG 2025-09 unverdicted novelty 6.0 of 10

    LoFT uses parameter-efficient fine-tuning of foundation models for long-tailed semi-supervised learning, supported by proofs that this reduces hypothesis complexity to minimize balanced posterior error and compresses ...

  15. Uncertainty-Aware Likelihood Ratio Estimation for Pixel-Wise Out-of-Distribution Detection

    cs.CV 2025-08 conditional novelty 6.0 of 10

    An evidential classifier trained on synthetic outliers reduces the average false-positive rate to 2.5% for pixel-wise out-of-distribution detection in road-scene segmentation.

  16. Text-ADBench: Text Anomaly Detection Benchmark Based on LLM Embeddings

    cs.CL 2025-07 conditional novelty 6.0 of 10

    LLM embeddings improve text anomaly detection, shallow detectors match deep ones only under oracle embedding selection, and AUROC matrices are low-rank enough to support fast model evaluation.

  17. SODA: Out-of-Distribution Detection in Domain-Shifted Point Clouds via Neighborhood Propagation

    cs.CV 2025-06 conditional novelty 6.0 of 10

    A training-free neighborhood score propagation method improves out-of-distribution detection for point clouds under synthetic-to-real domain shift.

  18. How to Use Graph Data in the Wild to Help Graph Anomaly Detection?

    cs.LG 2025-06 conditional novelty 6.0 of 10

    Wild-GAD selects relevant and diverse external graphs via a target-trained model and trains the detector on them, reporting large accuracy gains over baselines.

  19. TerraIncognita: A Dynamic Benchmark for Species Discovery Using Frontier Models

    cs.CV 2025-05 conditional novelty 6.0 of 10

    A new dynamic benchmark using real field-collected images of rare insects shows frontier vision-language models are strong at coarse taxonomic classification but nearly fail at species-level identification and are inc...

  20. LoD: Loss-difference OOD Detection by Intentionally Label-Noisifying Unlabeled Wild Data

    cs.LG 2025-05 conditional novelty 6.0 of 10

    A threshold-free OOD detection method that labels wild data as an extra class and clusters per-sample training losses to separate in-distribution from out-of-distribution data.

  21. Graph Synthetic Out-of-Distribution Exposure with Large Language Models

    cs.LG 2025-04 conditional novelty 6.0 of 10

    LLM-identified or LLM-generated pseudo-OOD nodes used as exposure data during GNN training improve node-level OOD detection on text-attributed graphs without real OOD labels.

  22. OOD Detection with immature Models

    cs.LG 2025-02 conditional novelty 6.0 of 10

    Partially trained GLOW models match or outperform fully trained models for out-of-distribution image detection when scored by layer-wise gradient norms.

  23. HEM: a margin-based loss for visual categorisation tasks

    cs.LG 2025-01 conditional novelty 6.0 of 10

    A new margin-based loss, HEM, trains image classifiers that are more robust to unknown and adversarial inputs and better at continual learning and segmentation than cross-entropy-trained models.

  24. Soft Checksums to Flag Untrustworthy Machine Learning Surrogate Predictions and Application to Atomic Physics Simulations

    cs.LG 2024-12 conditional novelty 6.0 of 10

    A neural network trained with an extra checksum output can flag out-of-distribution predictions by measuring how strongly its own outputs violate the checksum relation.

  25. Theoretical Grounding of Out-Of-Distribution Detection With Reinforcement Learning Optimizer

    cs.CV 2026-06 unverdicted novelty 5.0 of 10

    Develops an RL-augmented gradient descent optimizer for dynamic OOD detection together with a temporal error decomposition framework comparing it to standard GD.

  26. TaskFusion: Continual Anomaly Detection for Heterogeneous Tabular Data

    cs.LG 2026-06 unverdicted novelty 5.0 of 10

    TaskFusion combines AGF feature mapping, cross-task augmentation, and distilled replay for continual anomaly detection on heterogeneous tabular data, reporting gains over baselines on 21 datasets.

  27. Dual Feature Decoupling for Fine-Grained OOD Detection

    cs.CV 2026-06 unverdicted novelty 5.0 of 10

    DFDNet disentangles content from style via dual modules to boost fine-grained OOD detection performance on multiple datasets.

  28. Holistic Reliability Propagation: Decoupling Annotation and Prediction for Robust Noisy-Label

    cs.CV 2026-05 unverdicted novelty 5.0 of 10

    HRP decouples annotation reliability (alpha) and pseudo-label reliability (beta) via bilevel meta-learning and routes them to distinct objectives in reliability-aware Mixup and contrastive learning for improved noisy-...

  29. Real-time Anomaly Detection for Liquid Argon Time Projection Chambers

    hep-ex 2025-09 conditional novelty 5.0 of 10

    An autoencoder trained on MicroBooNE LArTPC data, compressed via knowledge distillation, flags high-multiplicity track segments with ROC-AUC 0.93 on 864x64 inputs, while the FPGA-suitable 18x16 model is near random.

  30. DCV-ROOD Evaluation Framework: Dual Cross-Validation for Robust Out-of-Distribution Detection

    cs.LG 2025-09 conditional novelty 5.0 of 10

    DCV-ROOD is a dual cross-validation framework for OOD detection that splits ID data by stratified folds and OOD data by class groups, reproducing benchmark statistical comparisons at lower cost.

  31. Knowledge Regularized Negative Feature Tuning of Vision-Language Models for Out-of-Distribution Detection

    cs.CV 2025-07 conditional novelty 5.0 of 10

    KR-NFT tunes CLIP text features with image-conditioned scaling and shifting plus a knowledge regularization loss, improving OOD detection on base and unseen classes without forgetting pre-trained knowledge.

  32. WeedNet: A Foundation Model-Based Global-to-Local AI Approach for Real-Time Weed Species Identification and Classification

    cs.CV 2025-05 conditional novelty 5.0 of 10

    An AI model trained on citizen science images identifies over 1,500 weed species globally and fine-tunes to regional weed communities with high accuracy.

  33. Are vision language models robust to uncertain inputs?

    cs.CV 2025-05 conditional novelty 5.0 of 10

    Prompting VLMs to say "unknown" on ambiguous inputs substantially improves classification reliability on natural images, and caption diversity tracks this abstention behavior, though the mechanism fails on specialized...

  34. Open-set Anomaly Segmentation in Complex Scenarios

    cs.CV 2025-04 conditional novelty 5.0 of 10

    An adverse-weather anomaly segmentation benchmark shows current models fail badly, and a diffusion-plus-energy-entropy training method boosts their robustness.

  35. Bounded and Uniform Energy-based Out-of-distribution Detection for Graphs

    cs.LG 2025-04 conditional novelty 5.0 of 10

    NODESAFE adds two logit variance penalties to GNN training, cutting FPR95 for node-level OOD detection dramatically compared with GNNSAFE on citation and social-network graphs.

  36. Mitigating Spurious Negative Pairs for Robust Industrial Anomaly Detection

    cs.CV 2025-01 conditional novelty 5.0 of 10

    A contrastive anomaly detector trained on pseudo-anomalies and opposite-pair repulsion raises average robust AUROC under PGD-1000 from 39.7% (best prior) to 65.8%.

  37. A Unified Plug-and-Play Framework for Effective Data Denoising and Robust Abstention

    cs.LG 2020-09 conditional novelty 5.0 of 10

    A model-agnostic framework filters noisy training samples and abstains uncertain test samples by measuring distance to class centroids in DNN feature space, matching or beating DAC and SelectiveNet on several benchmarks.

  38. Out-of-Distribution Detection Using Neural Rendering Generative Models

    cs.LG 2019-07 unverdicted novelty 5.0 of 10

    NRM enables OoD detection by joint latent likelihood, assigning lower values to SVHN than CIFAR-10 (unlike VAEs/flows) and consistent across other OoD sets.

  39. At the Edge of Understanding: Sparse Autoencoders Trace The Limits of Transformer Generalization

    cs.LG 2026-06 unverdicted novelty 4.0 of 10

    Sparse autoencoders show OOD prompts increase fallacious concept activation in transformers, offering a mechanistic measure of shift and a path to robust fine-tuning.

  40. DMDSC: A Dynamic-Margin Deep Simplex Classifier for Open-Set Recognition on Medical Image Datasets

    cs.CV 2026-05 unverdicted novelty 4.0 of 10

    DMDSC adapts simplex-classifier margins dynamically according to label frequency to tighten clustering on rare medical classes and improve open-set rejection on imbalanced imaging datasets.

  41. Out-of-distribution data supervision towards biomedical semantic segmentation

    cs.CV 2025-07 reject novelty 4.0 of 10

    Med-OoD adds background-only 'OOD' patches from the ID dataset as negative samples with zero-mask Dice loss, claiming modest gains on Lizard but an internally inconsistent 76.1% mIoU for the OOD-only case.

  42. Adaptive Deviation Learning for Visual Anomaly Detection with Data Contamination

    cs.CV 2024-11 reject novelty 4.0 of 10

    Adaptive Deviation Learning combines a soft-label deviation loss with instance reweighting to improve visual anomaly detection when the training set contains unlabeled anomalies.

  43. Detecting Hope, Hate, and Emotion in Arabic Textual Speech and Multi-modal Memes Using Large Language Models

    cs.CL 2025-08 unverdicted novelty 3.0 of 10

    The submission cannot be reviewed as a coherent paper: its abstract and full text are two different papers, so the abstract's claims have no supporting body.

  44. From Pixel to Mask: A Survey of Out-of-Distribution Segmentation

    cs.CV 2025-08 conditional novelty 3.0 of 10

    A survey categorizing out-of-distribution segmentation methods for autonomous driving into test-time, outlier-exposure, reconstruction, and powerful-model families.

  45. Handling Out-of-Distribution Data: A Survey

    cs.LG 2025-07 conditional novelty 3.0 of 10

    A survey that organizes covariate and semantic shift handling methods into one taxonomy and argues for unified models, while contributing no new experiments or benchmark evaluation.

  46. RODEO: Robust Outlier Detection via Exposing Adaptive Out-of-Distribution Samples

    cs.CV 2025-01

Pith tools