Reproduce “ExCal: Optimization-Free Multivariate Time Series Anomaly Detection via Excess Calibration” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Kadir-Kaan Özer, René Ebeling, Markus Enzweiler
ExCal detects anomalies in telemetry by combining tests of normal behavior derived from a training prefix. It calibrates each test’s positive excess using prefix statistics before fusing the scores. Across 180 series, ExCal ranks second among 24 methods on AUC-PR and VUS-PR, with no significant adjusted difference from AxonAD or StreamVAE; its scoring is batch and transductive.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.