Research map / Label-efficient and robust learning / OOD and anomaly detection
Transfer Learning for Anomaly Detection through Localized and Unsupervised Instance Selection
Vercruyssen Vincent, Meert Wannes, Davis Jesse
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| use SSKNNO instead of KNN as target classifier (ablation) | CORAL transfer · Anomaly transfer benchmark | average AUROC: 0.666 -> 0.735 | Improved |
| use SSKNNO instead of KNN as target classifier (ablation) | TransferAll · Anomaly transfer benchmark | average AUROC: 0.649 -> 0.727 | Improved |
| use SSKNNO instead of KNN as target classifier (ablation) | GFK transfer · Anomaly transfer benchmark | average AUROC: 0.642 -> 0.705 | Improved |
| use SSKNNO instead of KNN as target classifier (ablation) | CBIT transfer · Anomaly transfer benchmark | average AUROC: 0.623 -> 0.713 | Improved |
| use SSKNNO instead of KNN as target classifier (ablation) | TCA transfer · Anomaly transfer benchmark | average AUROC: 0.656 -> 0.533 | Hurt |
| replace global domain alignment/transfer-all with localized instance selection (LocIT transfer) under same SSKNNO classifier (ablation) | SSKNNO · Anomaly transfer benchmark | average AUROC: 0.735 -> 0.762 | Improved |
| add transferred labeled source instances (LocIT) over unsupervised detection | anomaly detector · Anomaly transfer benchmark | average AUROC: 0.705 -> 0.762 | Improved |
| increase neighborhood size psi to >= 20 (ablation) | LocIT · Anomaly transfer benchmark | AUROC | Improved |
| increase fraction of labeled source instances (ablation) | LocIT · Anomaly transfer benchmark | AUROC | Improved |
| LocIT transfer vs KNNO on real water usage data | LocIT · Water usage, 6 store pairs | average AUROC: 0.754 -> 0.779 | Mixed |
Closest papers by meaning
- Transfer Anomaly Detection by Inferring Latent Domain Representations (NeurIPS 2019 · similarity 0.86)
- Tight Rates in Supervised Outlier Transfer Learning (ICLR 2024 · similarity 0.84)
- Normal-Abnormal Guided Generalist Anomaly Detection (NeurIPS 2025 · similarity 0.83)
- A Comprehensive Augmentation Framework for Anomaly Detection (AAAI 2024 · similarity 0.83)
- We Need to Rethink Benchmarking in Anomaly Detection (NeurIPS 2026 · similarity 0.83)
- Unsupervised Anomaly Detection Algorithms on Real-world Data: How Many Do We Need? (NeurIPS 2024 · similarity 0.82)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
