Research map / Label-efficient and robust learning / OOD and anomaly detection
Deep one-class and tabular anomaly detection
Applies one-class objectives, self-supervision and reinforcement learning to tabular and general data.
50 accepted papers at ICML, NeurIPS, ICLR, CVPR and AAAI (2018–2026) · an approach in OOD and anomaly detection, keyword cluster “anomaly detection · normal · anomalies”
State History Tried Checklist Preprints Who Export
Where it stands
Not enough benchmark data Fewer than four papers report the same benchmark, so there is no record to follow.
15 papers in 2023–24 and 20 in 2025–26 (at ICML, NeurIPS, ICLR, CVPR and AAAI); its share of the topic held steady, from 7.1% to 7.7%.
How this is decided
A benchmark is tracked when at least four papers of this approach report their own method on the same dataset and metric in at least two different years. Its record is the best number reported up to 2024 compared with the best reported in 2025–26: still improving means at least 1% better, flat means at most 0.2% better with at least two papers from 2025–26 trying. Numbers on a different scale from the rest (a fraction among percentages) and far outliers are left out. Results come from different papers with different setups, so they are signals, not head-to-head comparisons. The share of papers is counted at ICML, NeurIPS, ICLR, CVPR and AAAI.
How it got here
Most cited paper of each year
- 2018 · Deep One-Class Classification (ICML 2018 · 847 citations)
- 2019 · Anomaly Detection With Multiple-Hypotheses Predictions (ICML 2019 · 36 citations)
- 2020 · Old Is Gold: Redefining the Adversarially Learned One-Class Classifier Training Paradigm (CVPR 2020 · 190 citations)
- 2021 · Explainable Deep One-Class Classification (ICLR 2021 · 102 citations)
- 2022 · Self-Supervised Predictive Convolutional Attentive Block for Anomaly Detection (CVPR 2022 · 288 citations)
- 2023 · One-for-All: Proposal Masked Cross-Class Anomaly Detection (AAAI 2023 · 31 citations)
- 2024 · PromptAD: Learning Prompts with only Normal Samples for Few-Shot Anomaly Detection (CVPR 2024 · 135 citations)
- 2025 · Distribution Prototype Diffusion Learning for Open-set Supervised Anomaly Detection (CVPR 2025 · 7 citations)
- 2026 · Robust Detection of Synthetic Tabular Data Under Schema Variability (AAAI 2026 · 1 citation)
Earliest papers here
- Deep One-Class Classification (ICML 2018 · 847 citations)
- Adversarially Learned One-Class Classifier for Novelty Detection (CVPR 2018 · 769 citations)
- Multi-Cell Detection and Classification Using a Generative Convolutional Model (CVPR 2018 · 6 citations)
What has been tried
34 papers here report 298 measured changes (a component added, removed or swapped, compared with a baseline): helped 71%, no clear effect 5%, hurt 19%, mixed 4%. Papers mostly publish changes that helped, so "helped" is over-represented.
Ablations in recent papers
| What was changed | Measured on | Result | Effect | Paper |
|---|---|---|---|---|
| Add semantic concatenation to conventional prompt learning | MVTec 1-shot · image-level AUROC | 81.5 -> 90.4 | Helped | PromptAD: Learning Prompts with only Normal Samples for Few-Shot Anomaly Detection (CVPR 2024 · 135 citations) |
| add source-domain pre-training before target-domain tuning | BGL · F1 | Helped | LogFormer: A Pre-train and Tuning Pipeline for Log Anomaly Detection (AAAI 2024 · 70 citations) | |
| add adaptive weights eᵂᵢⱼ | CIFAR-10 · AUC (%) | 68.5 -> 74.5 | Helped | Dense Projection for Anomaly Detection (AAAI 2024 · 11 citations) |
| Add distribution prototype learning (DPL) | Industrial anomaly datasets, general and hard settings · AUC | Helped | Distribution Prototype Diffusion Learning for Open-set Supervised Anomaly Detection (CVPR 2025 · 7 citations) | |
| Use two distinct subsets with two-head self-attention instead of one attention head and one subset | Thyroid · AUC-PR | 72.6^{±1.6} -> 89.4^{±0.7} | Helped | Disentangling Tabular Data Towards Better One-Class Anomaly Detection (AAAI 2025 · 5 citations) |
| use Gaussian noise level around 1.0 | OCC · AUC | Helped | Unsupervised Anomaly Detection for Tabular Data Using Deep Noise Evaluation (AAAI 2025 · 4 citations) | |
| Add one abnormal reference and process normal-abnormal residuals with NAGL rather than using normal-reference NN only | MVTecAD · Image AUROC | 93.2 -> 95.8; +2.6 | Helped | Normal-Abnormal Guided Generalist Anomaly Detection (NeurIPS 2025 · 2 citations) |
| combine feature-feature and sample-sample dependencies | Ablation study · F1 | 57.4 -> 68.8; 57.5 -> 68.8 | Helped | Beyond Individual Input for Deep Anomaly Detection on Tabular Data (ICML 2024 · 1 citation) |
| add channel-level reference | Clinical · F2 | 51.20 -> 55.93 | Helped | DMNet: Self-comparison Driven Model for Subject-independent Seizure Detection (NeurIPS 2024 · 1 citation) |
| add contrastive reverse distillation to RD4AD | RSNA · AUC | 84.29 -> 89.87 | Helped | Scale-Aware Contrastive Reverse Distillation for Unsupervised Medical Anomaly Detection (ICLR 2025) |
| replace a single large detector with MoME | ADBench (40/60% train/test) · AUC-ROC | 0.91 -> 0.93 | Helped | Swift Hydra: Self-Reinforcing Generative Framework for Anomaly Detection with Multiple Mamba Models (ICLR 2025) |
| add OT prototype matching | MVTec-AD · image-level AUROC | 68.4 -> 93.5 | Helped | Vague Prototype-Oriented Diffusion Model for Multi-Class Anomaly Detection (ICML 2024) |
What reviewers will expect
Datasets and baselines named in the 13 papers of this approach from 2025–2026 that have a full-text summary. Leaving out the common ones is a frequent review question.
Datasets
- MVTec AD · 2 of 13 papers
- ADBench · 2 of 13 papers
Baselines
- DeepSVDD · 3 of 13 papers
- IForest · 2 of 13 papers
- LOF · 2 of 13 papers
- ECOD · 2 of 13 papers
- NeuTraLAD · 2 of 13 papers
- AE · 2 of 13 papers
- KNN · 2 of 13 papers
- PatchCore · 2 of 13 papers
Recent preprints
Loading recent arXiv preprints…
Who publishes on it
Institutions on the most papers of this approach since 2023 (of 35 papers; from OpenAlex and the papers' affiliations).
- The Chinese University of Hong Kong · 3 papers
- Xi’an Jiaotong-Liverpool University · 3 papers
- University of Liverpool · 3 papers
- Alibaba Group · 2 papers
Use it
Download BibTeX of 35 key papersOpen the topic in the interactive map
Get alerts. Describe your idea in scime atlas and save it (free scime account): you get a weekly email when new preprints land close to it.
Related approaches
Same keyword cluster
- Unified multi-class anomaly detection (86 papers)
- Anomaly detection with contaminated data (64 papers)
Elsewhere in the topic
- Provable and feature-based OOD detection (122 papers)
- Spurious-aware OOD detection (109 papers)
- Out-of-distribution generalization theory (71 papers)
- Uncertainty and misclassification detection (56 papers)
- Robust fine-tuning for OOD generalization (56 papers)
Benchmark numbers, ablations and baseline numbers are extracted by AI from each paper's full text and can be wrong; the paper is the authority. Data: results index of 2026-11-02, citation counts from OpenAlex (2026-10-09).