Statistical Analysis of using the Shapley Value for Sensor Anomaly Localization with Accurate Classifiers
arXiv:2606.00867v1 Announce Type: cross Abstract: Recent publications have suggested using the Shap- ley value for sensor anomaly/attack localization. We study the performance of such an approach by using mathematically de- fined optimum binary classifiers in the Shapley value calculation. To judge localization performance, we study the ability of the Shapley value of a given sensor observation to determine if that observation is anomalous. First, we prove that for cases with independent sensor observations, an optimized anomaly test using the Shapley value is equivalent to an optimized lower-
This paper presents foundational research in the application of the Shapley value for sensor anomaly detection, a topic that has seen increased attention in recent academic discourse.
For a sophisticated reader, it refines understanding of anomaly detection methods, particularly relevant for AI and embedded systems, but does not represent an immediate practical breakthrough.
This research provides a theoretical validation and optimization framework for using Shapley values in sensor anomaly localization, enhancing the fidelity of diagnostic AI systems in the long run.
Improved theoretical understanding of anomaly detection using Shapley values.
Potential for more robust AI systems capable of self-diagnosing sensor issues.
Enhanced reliability and safety in autonomous systems relying on sensor data, albeit in a distant future.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.LG