![[2602.12592] Power Interpretable Causal ODE Networks: A Unified Model for Explainable Anomaly Detection and Root Cause Analysis in Power Systems](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.12592] Power Interpretable Causal ODE Networks: A Unified Model for Explainable Anomaly Detection and Root Cause Analysis in Power Systems
Summary
The paper presents Power Interpretable Causal ODE Networks (PICODE), a novel model for explainable anomaly detection and root cause analysis in power systems, enhancing interpretability and performance in identifying anomalies.
Why It Matters
As cyber-physical systems like power grids become increasingly complex, the need for transparent anomaly detection methods is critical. This research addresses the limitations of existing models by providing explanations for detected anomalies, which is essential for safety and operational efficiency in power systems.
Key Takeaways
- PICODE Networks improve anomaly detection with enhanced interpretability.
- The model reduces reliance on labeled data and external causal graphs.
- Theoretical results link anomaly function shapes to causal graph weight changes.
- Experimental results show competitive performance in power systems.
- The approach supports root cause localization and anomaly type classification.
Computer Science > Machine Learning arXiv:2602.12592 (cs) [Submitted on 13 Feb 2026] Title:Power Interpretable Causal ODE Networks: A Unified Model for Explainable Anomaly Detection and Root Cause Analysis in Power Systems Authors:Yue Sun, Likai Wang, Rick S. Blum, Parv Venkitasubramaniam View a PDF of the paper titled Power Interpretable Causal ODE Networks: A Unified Model for Explainable Anomaly Detection and Root Cause Analysis in Power Systems, by Yue Sun and 3 other authors View PDF HTML (experimental) Abstract:Anomaly detection and root cause analysis (RCA) are critical for ensuring the safety and resilience of cyber-physical systems such as power grids. However, existing machine learning models for time series anomaly detection often operate as black boxes, offering only binary outputs without any explanation, such as identifying anomaly type and origin. To address this challenge, we propose Power Interpretable Causality Ordinary Differential Equation (PICODE) Networks, a unified, causality-informed architecture that jointly performs anomaly detection along with the explanation why it is detected as an anomaly, including root cause localization, anomaly type classification, and anomaly shape characterization. Experimental results in power systems demonstrate that PICODE achieves competitive detection performance while offering improved interpretability and reduced reliance on labeled data or external causal graphs. We provide theoretical results demonstrating the a...