Forecasting Anomaly Precursors via Uncertainty-Aware Time-Series Ensembles

Detecting anomalies in time-series data is critical in domains such as industrial operations, finance, and cybersecurity, where early identification of abnormal patterns is essential for ensuring system reliability and enabling preventive maintenance. However, most existing methods are reactive: they detect anomalies only after they occur and lack the capability to provide proactive early warning signals. In this paper, we propose FATE (Forecasting Anomalies with Time-series Ensembles), a novel unsupervised framework for detecting Precursors-of-Anomaly (PoA) by quantifying predictive uncertainty from a diverse ensemble of time-series forecasting models. Unlike prior approaches that rely on reconstruction errors or require ground-truth labels, FATE anticipates future values and leverages ensemble disagreement to signal early signs of potential anomalies without access to target values at inference time. To rigorously evaluate PoA detection, we introduce Precursor Time-series Aware Precision and Recall (PTaPR), a new metric that extends the traditional Time-series Aware Precision and Recall (TaPR) by jointly assessing segment-level accuracy, within-segment coverage, and temporal promptness of early predictions. This enables a more holistic assessment of early warning capabilities that existing metrics overlook. Experiments on five real-world benchmark datasets show that FATE achieves an average improvement of 19.9 percentage points in PTaPR AUC and 20.02 percentage points in early detection F1 score, outperforming baselines while requiring no anomaly labels. These results demonstrate the effectiveness and practicality of FATE for real-time unsupervised early warning in complex time-series environments.

💡 Research Summary

The paper tackles the problem of early warning for anomalies in multivariate time‑series, a task the authors term Precursor‑of‑Anomaly (PoA) detection. Unlike conventional anomaly detection, which reacts after an abnormal event has already manifested, PoA detection aims to signal the impending anomaly before it occurs, thereby enabling preventive actions in safety‑critical domains such as manufacturing, energy, and cybersecurity.

To achieve this, the authors propose FATE (Forecasting Anomalies with Time‑series Ensembles), an unsupervised framework that leverages the predictive uncertainty of an ensemble of diverse forecasting models. The core idea is simple yet powerful: when the input sequence contains latent symptoms of an upcoming fault, different forecasting models will disagree more strongly about the future values. This disagreement is quantified as the variance (or more generally, any dispersion measure) of the predicted values at each future time step. A high variance is interpreted as a high likelihood of a PoA.

Framework details

1. Sliding‑window formulation – The raw series (T={x_1,\dots,x_T}) is segmented into overlapping windows of length (L_x). Each window (X_t) serves as the input to all models in the ensemble.
2. Ensemble of forecasters – The ensemble comprises (N) state‑of‑the‑art predictors (e.g., Transformer‑based models, LSTM, N‑Beats, etc.). Each model is trained independently to map (X_t) to a multi‑step forecast (Y_t^{(i)}={y_{t,1}^{(i)},\dots,y_{t,L_y}^{(i)}}).
3. Uncertainty estimation – For each future horizon (\tau\in