Research map / Label-efficient and robust learning / OOD and anomaly detection
Interpreting Unsupervised Anomaly Detection in Security via Rule Extraction
Ruoyu Li, Qing Li, Yu Zhang, Dan Zhao, Yong Jiang, Yong Yang
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 IC-Tree with K-Means (k=10) (ablation) | IC-Tree + CBE · CIC-IDS2017 | fidelity: 0.9856 -> 0.9731 | Hurt |
| replace IC-Tree with K-Means (k=5) (ablation) | IC-Tree + CBE · CIC-IDS2017 | fidelity: 0.9856 -> 0.9735 | Hurt |
| use only CBE without IC-Tree (ablation) | IC-Tree + CBE · CIC-IDS2017 | fidelity: 0.9856 -> 0.9735 | Hurt |
| use hypercubes as rules instead of CBE (ablation) | IC-Tree + CBE · CIC-IDS2017 | fidelity: 0.9856 -> 0.9652 | Hurt |
| use hypercubes as rules instead of CBE (ablation) | IC-Tree + CBE · TON-IoT | fidelity: 0.9840 -> 0.0647 | Hurt |
| replace IC-Tree with K-Means (k=10) (ablation) | IC-Tree + CBE · TON-IoT | fidelity: 0.9840 -> 0.9802 | Hurt |
| use only CBE without IC-Tree (ablation) | IC-Tree + CBE · TON-IoT | fidelity: 0.9840 -> 0.9784 | Hurt |
| add 10% random noise to training data | Ours on iForest · random noise | fidelity: 0.9876 -> 0.5148 | Hurt |
| add 10% random noise to training data | Ours on AE · random noise | fidelity: 0.9829 -> 0.9739 | No clear effect |
| add 10% mislabeled attack data to training | Ours on AE · mislabeled noise | fidelity: 0.9997 -> 0.9987 | No clear effect |
| use rule extraction Ours instead of EGDT | rule extraction · CIC-IDS2017 AE | fidelity: 0.533 -> 0.9835 | Improved |
| use Ours instead of UAD rule extraction | rule extraction · TON-IoT AE | fidelity: 0.1499 -> 0.9996 | Improved |
Closest papers by meaning
- Dissect Black Box: Interpreting for Rule-Based Explanations in Unsupervised Anomaly Detection (NeurIPS 2024 · similarity 0.92)
- Towards a Unified Framework of Clustering-based Anomaly Detection (ICML 2025 · similarity 0.87)
- Unsupervised Anomaly Detection Algorithms on Real-world Data: How Many Do We Need? (NeurIPS 2024 · similarity 0.85)
- Unifying Unsupervised Graph-Level Anomaly Detection and Out-of-Distribution Detection: A Benchmark (ICLR 2025 · similarity 0.84)
- Unsupervised Anomaly Detection in The Presence of Missing Values (NeurIPS 2024 · similarity 0.83)
- SEAD: Unsupervised Ensemble of Streaming Anomaly Detectors (ICML 2025 · 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.
