Collaborative system improves detection of streaming data anomalies
Collaborative Streaming Anomaly Detection with Interactive Explanations and Ensemble Consensus
Machine Learning
Summary
Detecting unusual patterns in fast-moving data streams is hard because machines and humans see information differently. The authors describe a new system where different automated detectors’ opinions are combined, and human analysts can interactively adjust and understand these guesses. This system also builds simple explanations that help people see why something looks suspicious. They tested it on industrial data showing it can reliably find rare problems while letting humans guide the process.
What this means in practice
- •For industrial system operators: Monitor real-time sensor streams to pinpoint rare faults with collaborative human-machine detection and interactive feedback.
- •For network security analysts: Combine multiple anomaly detectors on network traffic streams and let analysts refine alerts for better threat identification.
Tested on one dataset.
Authors
Diogo Risca, Afonso Lourenço, Ricardo Martins, Goreti Marreiros
Abstract
We present a collaborative streaming anomaly detection system for high-speed data streams that explicitly integrates human analysts into the decision loop. The system combines heterogeneous detectors and aggregates their outputs through a normalization-based weighted consensus, complemented by artifact-aware rules to stabilize anomaly scoring under deployment. To improve interpretability, it derives surrogate models that approximate the ensemble consensus and expose human-readable sensor conditions associated with anomalous behavior. Analysts can actively intervene by reviewing anomaly episodes, adjusting consensus behavior, and refining surrogate rules used for anomaly prediction, producing a human-adjusted ensemble. We evaluate the approach on an industrial stream with 260\,000 events and 3 anomalous episodes, showing robust detection and actionable human-AI interaction.