Research map / Label-efficient and robust learning / OOD and anomaly detection
OnlineBootKNN: An Unsupervised Framework for Detecting Anomalies in Spectral Data Streams
Nicolas Rojas Varela, Julien Ah-Pine, Engelbert Mephu Nguifo
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 the OBKNN framework (z-normalization, Poisson bootstrapped chunks, z-score test) instead of plain SWKNN (ablation) | sliding-window KNN · Scenarios A4-A6 | AUC-PR: 0.918 -> 0.933 | Improved |
| increase k from 1 to 10 or 50 in SWKNN (ablation) | SWKNN · Scenarios A4-A6 | AUC-PR: 0.918 -> 0.746 -> 0.517 | Hurt |
| OBKNN versus KitNet overall | OBKNN · Spectral leak streams A1-A9 | AUC-PR: 0.935 vs 0.933 | No clear effect |
| OBKNN has far lower run-to-run variance than other methods | OBKNN · Spectral leak streams A1-A9 | AUC-PR std: std ~1e-4 vs ~1e-2 for most others | Improved |
| autoencoder ensemble learning from whole history (KitNet) on short leaks | KitNet · Scenarios A1, A4, A7 | AUC-PR: A4 0.786 vs A2 0.998 | Hurt |
| retrain trees per window (IFASD) on short versus long leaks | IFASD · Scenarios A1/A4/A7 vs others | AUC-PR: A1 0.975, A3 0.565 | Mixed |
| use OBKNN instead of KitNet in compute cost | OBKNN · Scenario A6 | processing time (s): 204.01 -> 1557.89 | Hurt |
Closest papers by meaning
- ScatterAD: Temporal-Topological Scattering Mechanism for Time Series Anomaly Detection (NeurIPS 2025 · similarity 0.83)
- A Unified Latent Schrodinger Bridge Diffusion Model for Unsupervised Anomaly Detection and Localization (CVPR 2025 · similarity 0.83)
- AutoSciDACT: Automated Scientific Discovery through Contrastive Embedding and Hypothesis Testing (NeurIPS 2025 · similarity 0.83)
- Multi-Sensor Object Anomaly Detection: Unifying Appearance, Geometry, and Internal Properties (CVPR 2025 · similarity 0.82)
- Rayleigh Quotient Graph Neural Networks for Graph-level Anomaly Detection (ICLR 2024 · similarity 0.82)
- Solving Spectrum Unmixing as a Multi-Task Bayesian Inverse Problem with Latent Factors for Endmember Variability (AAAI 2024 · similarity 0.81)
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.
