Research map / Time series / Anomaly detection
Cluster-Aware Causal Mixer for Online Anomaly Detection in Multivariate Time Series
Md Mahmuddun Nabi Murad, Yasin Yilmaz
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 point-based scoring with sequential evidence accumulation plus boundary correction (ablation) | CCM-TAD · PSM | best-F1: 0.657 -> 0.716 | Improved |
| replace point-based scoring with sequential evidence accumulation plus boundary correction (ablation) | CCM-TAD · MSL | best-F1: 0.557 -> 0.638 | Improved |
| add boundary correction to sequential scoring (ablation) | CCM-TAD · SWaT | best-F1: 0.879 -> 0.883 | No clear effect |
| add channel clustering (Case 5 vs Case 10) (ablation) | CCM-TAD · PSM | best-F1: 0.703 -> 0.716 | Improved |
| add channel clustering (Case 5 vs Case 10) (ablation) | CCM-TAD · SWaT | best-F1: 0.865 -> 0.883 | Improved |
| use causal temporal mixer instead of non-causal mixer (with clustering) (ablation) | CCM-TAD · PSM | best-F1: 0.652 -> 0.716 | Improved |
| add non-causal temporal mixer (vs none, with clustering) (ablation) | CCM-TAD · PSM | best-F1: 0.657 -> 0.652 | No clear effect |
| add embedding mixer (causal mixer, with clustering) (ablation) | CCM-TAD · PSM | best-F1: 0.681 -> 0.716 | Improved |
| add channel clustering to embedding (ablation) | CCM-TAD · SWaT | NN-sep: 65.63 -> 91.30 | Improved |
| add channel clustering to embedding (ablation) | CCM-TAD · PSM | NN-sep: 66.74 -> 66.78 | No clear effect |
| use spectral clustering on full correlation profile instead of K-Means on profiles (ablation) | CCM-TAD · PSM | best-F1: 0.673 -> 0.716 | Improved |
| use CCM-TAD instead of SensitiveHUE (Offline) | CCM-TAD · SWaT | best-F1: 0.911 -> 0.883 | Hurt |
Closest papers by meaning
- Structured Temporal Causality for Interpretable Multivariate Time Series Anomaly Detection (NeurIPS 2025 · similarity 0.90)
- CATCH: Channel-Aware Multivariate Time Series Anomaly Detection via Frequency Patching (ICLR 2025 · similarity 0.89)
- Causality-Aware Contrastive Learning for Robust Multivariate Time-Series Anomaly Detection (ICML 2025 · similarity 0.88)
- GCAD: Anomaly Detection in Multivariate Time Series from the Perspective of Granger Causality (AAAI 2025 · similarity 0.88)
- The Elephant in the Room: Towards A Reliable Time-Series Anomaly Detection Benchmark (NeurIPS 2024 · similarity 0.87)
- COMET: Codebook-based Online-adaptive Multi-scale Embedding for Time-series Anomaly Detection (NeurIPS 2026 · similarity 0.87)
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.
