Projects

Exact, fast computation of the VUS measure on CPU and GPU, plus adaptive per-anomaly buffering for time series anomaly detection evaluation.

Interactive web app to read anomaly detection benchmarks series by series, with a game where you mark anomalies against the best detector.

Python package for loading the TSB-AD benchmark (series, anomaly scores, metrics, runtimes, splits) with an interactive Streamlit explorer.

Project template to set up time series anomaly detection experiments quickly and consistently.

Code for the MSAD journal paper on model selection for time series anomaly detection.

Demo of model selection for time series anomaly detection.