Research map / Graph learning / Graph self-supervised learning
Midas: Microcluster-Based Detector of Anomalies in Edge Streams
Siddharth Bhatia, Bryan Hooi, Minji Yoon, Kijung Shin, Christos Faloutsos
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 streaming microcluster scoring with Count-Min Sketch counts and hypothesis testing instead of SedanSpot | Edge-stream anomaly detection · DARPA | AUC: 0.64 -> 0.91 | Improved |
| Use streaming microcluster scoring with Count-Min Sketch counts and hypothesis testing instead of SedanSpot | Edge-stream anomaly detection · DARPA | Average Precision Score: 0.75 -> 0.95 | Improved |
| Use Midas instead of SedanSpot | Edge-stream anomaly detection · DARPA | runtime: 0.13s vs. 84s | Improved |
| Add temporal decay and spatial node relations | Midas · DARPA | AUC: 0.91 -> 0.95 | Improved |
| Add temporal decay and spatial node relations | Midas · DARPA | Average Precision Score: 0.95 -> 0.97 | Improved |
| Add temporal decay and spatial node relations | Midas · DARPA | runtime: 0.13s -> 0.39s | Hurt |
| Use Midas instead of SedanSpot | Edge-stream anomaly detection · TwitterWorldCup | runtime: 27.58s -> 0.06s | Improved |
| Use Midas instead of SedanSpot | Edge-stream anomaly detection · TwitterSecurity | runtime: 40.71s -> 0.08s | Improved |
| Use Midas instead of SedanSpot | Edge-stream anomaly detection · DARPA | runtime: 83.66s -> 0.13s | Improved |
| Use relational temporal and spatial edge grouping | Midas-R · TwitterSecurity | microcluster detection | Improved |
Closest papers by meaning
- UniGAD: Unifying Multi-level Graph Anomaly Detection (NeurIPS 2024 · similarity 0.84)
- Graph Anomaly Detection as Dynamical Transport: Training-Free Scoring via Empirical Bayes (NeurIPS 2026 · similarity 0.83)
- Dissect Black Box: Interpreting for Rule-Based Explanations in Unsupervised Anomaly Detection (NeurIPS 2024 · similarity 0.83)
- UniFORM: Towards Unified Framework for Anomaly Detection on Graphs (AAAI 2025 · similarity 0.83)
- Revisiting Graph Contrastive Learning on Anomaly Detection: A Structural Imbalance Perspective (AAAI 2025 · similarity 0.82)
- Rethinking Reconstruction-based Graph-Level Anomaly Detection: Limitations and a Simple Remedy (NeurIPS 2024 · similarity 0.82)
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.
