Research map / Label-efficient and robust learning / OOD and anomaly detection
Memorize Early, Then Query: Inlier-Memorization-Guided Active Outlier Detection
Minseo Kang, Seunghwan Park, Dongha Kim
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 RD for the first round and MM for the remaining rounds instead of MM only (ablation) | IMBoost · ADBench (57 datasets) | average test AUC: 0.847 -> 0.840 | Hurt |
| use mixture-model decision-boundary querying (MM) instead of random or confidence-poles querying (ablation) | IMBoost · ADBench (57 datasets) | average test AUC | Improved |
| query confidence poles (CP) instead of random sampling (ablation) | IMBoost · ADBench (57 datasets) | average test AUC | Hurt |
| set lambda1 or lambda2 (inlier or outlier risk weight) to zero (ablation) | IMBoost · ADBench (57 datasets) | AUC | Hurt |
| include inlier risk from queried inliers when determining adaptive threshold tau (ablation) | IMBoost · ADBench (57 datasets) | AUC | Improved |
| vary posterior probability threshold alpha (ablation) | IMBoost · ADBench (57 datasets) | AUC | No clear effect |
| set warm-up length T1 to 40 (ablation) | IMBoost · ADBench (57 datasets) | AUC | Improved |
| use IMBoost instead of the second-best baseline XGBOD+RD | active outlier detection · ADBench (57 datasets) | average test AUC: margin of nearly 5% | Improved |
Closest papers by meaning
- ALTBI: Constructing Improved Outlier Detection Models via Optimization of Inlier-Memorization Effect (AAAI 2025 · similarity 0.94)
- ODIM: Outlier Detection via Likelihood of Under-Fitted Generative Models (ICML 2024 · similarity 0.91)
- E3 Outlier: a Self-Supervised Framework for Unsupervised Deep Outlier Detection (TPAMI 2022 · similarity 0.89)
- DOS: Diverse Outlier Sampling for Out-of-Distribution Detection (ICLR 2024 · similarity 0.87)
- RODEO: Robust Outlier Detection via Exposing Adaptive Out-of-Distribution Samples (ICML 2024 · similarity 0.87)
- How Does Unlabeled Data Provably Help Out-of-Distribution Detection? (ICLR 2024 · similarity 0.86)
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.
