New method improves anomaly detection in changing time-series data
MORA: Modeling Observed Changes for Drift-Robust Time-Series Anomaly Detection
Machine LearningArtificial Intelligence
Summary
Detecting unusual events in data that changes over time is tricky because sometimes normal changes can look like problems. The authors introduce MORA, a system that compares short-term and long-term views of data to figure out if a change is truly unusual or just part of a normal shift. MORA adjusts how it judges anomalies based on how well it can explain changes in context, without needing extra labels or real-time updates. This makes it better at spotting real problems even when the data changes over time.
What this means in practice
- •For industrial monitoring teams: Improve detection of genuine faults in sensor data that exhibits changing patterns over time without mistaking normal shifts for failures.
- •For financial data analysts: Better identify true anomalies in financial time-series affected by market shifts, reducing false alarms from normal distribution changes.
Authors
Xudong Mou, Tiejun Wang, Rui Wang, Hui Wang, Pin Liu, Tianyu Wo, Xudong Liu, Renyu Yang
Abstract
Time-series anomaly detection (TSAD) identifies deviations from patterns learned from historical data. In non-stationary settings, distribution drift and true anomalies can cause similar local changes, making it difficult to tell whether a deviation reflects abnormality or evolving context. Existing methods typically adapt to detected shifts or learn drift-insensitive representations, but do not resolve this ambiguity. We define this problem as \emph{temporal change disambiguation}: determining whether a local deviation is explained by broader temporal evolution. We introduce MORA, a drift-robust TSAD framework that reconstructs the same local target from paired short- and long-term views. The reconstruction gap measures contextual support for a local deviation, and a data-dependent correction mechanism conservatively adjusts the primary local anomaly score. Context can only reduce the score when it improves reconstruction of the same target. MORA needs neither drift annotations nor online adaptation. Experiments on four TSAD benchmarks show strong robustness to non-stationarity while preserving sensitivity to genuine anomalies.