Research map / Label-efficient and robust learning / OOD and anomaly detection
OneFlow: One-class flow for anomaly detection based on a minimal volume region
Łukasz Maziarka, Marek Śmieja, Marcin Sendera, Łukasz Struski, Jacek Tabor, Przemysław Spurek
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 |
|---|---|---|---|
| replace log-likelihood loss with minimal-volume Bernstein quantile loss | flow model (LL-Flow vs OneFlow) · KDDCUP | F1: .6650 -> .9702 | Improved |
| replace log-likelihood loss with minimal-volume loss | flow model · Thyroid | F1: .6989 -> .7634 | Improved |
| use OneFlow loss instead of Deep SVDD loss | one-class flow · Thyroid | F1: .6989 -> .7634 | Improved |
| allow non-constant Jacobian (general flow) instead of const-det flow (ablation) | OneFlow · Thyroid | F1: .7634 -> .7097 | Hurt |
| allow non-constant Jacobian (general flow) (ablation) | OneFlow · KDDCUP | F1: .9702 -> .9712 | No clear effect |
| replace NICE backbone with Real-NVP (ablation) | OneFlow · MNIST | AUC: .964 -> .683 | Hurt |
| replace NICE backbone with Glow (ablation) | OneFlow · MNIST | AUC: .964 -> .895 | Hurt |
| replace NICE backbone with Real-NVP (ablation) | OneFlow · Fashion-MNIST | AUC: .922 -> .769 | Hurt |
| replace NICE backbone with Glow (ablation) | OneFlow · Fashion-MNIST | AUC: .922 -> .775 | Hurt |
| use OneFlow instead of LL-Flow for OOD detection | flow model · Fashion-MNIST vs MNIST outliers | detected outliers (%): 89.59% -> 92.47% | Improved |
| use OneFlow instead of LL-Flow for boundary length | flow model · Doughnut 2D | boundary length: 27.205 -> 7.482 | Improved |
| increase alpha from 1% to 10% when training data contain anomalies (ablation) | OneFlow · MNIST/Fashion-MNIST with corrupted training | AUC | Improved |
Closest papers by meaning
- Mixture Prototype Flow Matching for Open-Set Supervised Anomaly Detection (ICML 2026 · similarity 0.85)
- Scalable, Explainable and Provably Robust Anomaly Detection with One-Step Flow Matching (NeurIPS 2025 · similarity 0.84)
- ResAD: A Simple Framework for Class Generalizable Anomaly Detection (NeurIPS 2024 · similarity 0.84)
- One-to-Normal: Anomaly Personalization for Few-shot Anomaly Detection (NeurIPS 2024 · similarity 0.84)
- PO3AD: Predicting Point Offsets toward Better 3D Point Cloud Anomaly Detection (CVPR 2025 · similarity 0.83)
- Towards Scalable 3D Anomaly Detection and Localization: A Benchmark via 3D Anomaly Synthesis and A Self-Supervised Learning Network (CVPR 2024 · similarity 0.83)
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.
