MIXAD: Memory-Induced Explainable Time Series Anomaly Detection
This addresses the need for interpretable anomaly detection in industrial applications, offering a novel method that balances detection performance with explainability.
The paper tackles the problem of interpretable anomaly detection in multivariate time series data by introducing MIXAD, which uses a memory network and spatiotemporal processing to detect anomalies through shifts in memory activation patterns, resulting in a 34.30% and 34.51% improvement in interpretability metrics over state-of-the-art baselines.
For modern industrial applications, accurately detecting and diagnosing anomalies in multivariate time series data is essential. Despite such need, most state-of-the-art methods often prioritize detection performance over model interpretability. Addressing this gap, we introduce MIXAD (Memory-Induced Explainable Time Series Anomaly Detection), a model designed for interpretable anomaly detection. MIXAD leverages a memory network alongside spatiotemporal processing units to understand the intricate dynamics and topological structures inherent in sensor relationships. We also introduce a novel anomaly scoring method that detects significant shifts in memory activation patterns during anomalies. Our approach not only ensures decent detection performance but also outperforms state-of-the-art baselines by 34.30% and 34.51% in interpretability metrics.