Causally-Constrained Probabilistic Forecasting for Time-Series Anomaly Detection

ArXi:2604.17998v1 Announce Type: new Anomaly detection in multivariate time series is a central challenge in industrial monitoring, as failures frequently arise from complex temporal dynamics and cross-sensor interactions. While recent deep learning models, including graph neural networks and Transformers, have nstrated strong empirical performance, most approaches remain primarily correlational and offer limited for causal interpretation and root-cause localization. This study