ADE at XKDD 2026 — ECML PKDD 2026 🇮🇹

I am happy to share that our paper “Interpretable Ensemble Anomaly Detection for Cybersecurity Time Series” has been accepted at the 8th ECML PKDD International Workshop on eXplainable Knowledge Discovery in Data Mining — XKDD 2026!

This work was developed together with Martina Cinquini, Riccardo Guidotti, and Emanuele Sciancalepore.

In the paper, we introduce AnomalyDetectorEnsemble (ADE), an unsupervised framework for interpretable anomaly detection in cybersecurity time series.

Cybersecurity anomaly detection often relies on individual detectors whose specific inductive biases allow them to identify only particular kinds of anomalous behaviour. At the same time, many anomaly detection systems provide an anomaly score without explaining why an observation or temporal interval has been identified as suspicious.

ADE addresses both aspects by combining multiple heterogeneous anomaly detection paradigms with explanation mechanisms that are directly connected to the logic of each detector.

Our experiments on real-world network telemetry show that combining these complementary detection strategies can achieve a better precision–recall trade-off than the individual components, while the integrated explanation layer provides additional temporal context that can support analysts during the inspection of detected alerts.

Looking forward to presenting our work at XKDD 2026 and joining the broader ECML PKDD 2026 community.

See you in Naples! 🇮🇹