Graph Attention Networks with Physical Constraints for Anomaly Detection
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866914263177625600 |
|---|---|
| author | Homaei, Mohammadhossein Khazrak, Iman Molano, Ruben Caro, Andres Avila, Mar |
| author_facet | Homaei, Mohammadhossein Khazrak, Iman Molano, Ruben Caro, Andres Avila, Mar |
| contents | Water distribution systems (WDSs) face increasing cyber-physical risks, which make reliable anomaly detection essential. Many data-driven models ignore network topology and are hard to interpret, while model-based ones depend strongly on parameter accuracy. This work proposes a hydraulic-aware graph attention network using normalized conservation law violations as features. It combines mass and energy balance residuals with graph attention and bidirectional LSTM to learn spatio-temporal patterns. A multi-scale module aggregates detection scores from node to network level. On the BATADAL dataset, it reaches $F1=0.979$, showing $3.3$pp gain and high robustness under $15\%$ parameter noise. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_12426 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Graph Attention Networks with Physical Constraints for Anomaly Detection Homaei, Mohammadhossein Khazrak, Iman Molano, Ruben Caro, Andres Avila, Mar Machine Learning Cryptography and Security Water distribution systems (WDSs) face increasing cyber-physical risks, which make reliable anomaly detection essential. Many data-driven models ignore network topology and are hard to interpret, while model-based ones depend strongly on parameter accuracy. This work proposes a hydraulic-aware graph attention network using normalized conservation law violations as features. It combines mass and energy balance residuals with graph attention and bidirectional LSTM to learn spatio-temporal patterns. A multi-scale module aggregates detection scores from node to network level. On the BATADAL dataset, it reaches $F1=0.979$, showing $3.3$pp gain and high robustness under $15\%$ parameter noise. |
| title | Graph Attention Networks with Physical Constraints for Anomaly Detection |
| topic | Machine Learning Cryptography and Security |
| url | https://arxiv.org/abs/2601.12426 |