UniMAP: Model-free detection of unclassified noise transients in LIGO-Virgo data using the Temporal Outlier Factor

TL;DR

UniMAP is a model-free, multi-window anomaly detection pipeline that effectively identifies diverse transient noise glitches in LIGO-Virgo gravitational wave data, improving detection of true signals.

Contribution

This work introduces UniMAP, a novel multi-window, model-free anomaly detection method leveraging the Temporal Outlier Factor for identifying diverse noise transients in gravitational wave data.

Findings

Successfully detected glitches in LIGO and Virgo data

Extended anomaly detection to arbitrary transient morphologies

Demonstrated effectiveness during third observing run

Abstract

Data from current gravitational wave detectors contains a high rate of transient noise (glitches) that can trigger false detections and obscure true astrophysical events. Existing noise-detection algorithms largely rely on model-based methods that may miss noise transients unwitnessed by auxiliary sensors or with exotic morphologies. We propose the Unicorn Multi-window Anomaly-detection Pipeline (UniMAP): a model-free algorithm to identify and characterize transient noise leveraging the Temporal Outlier Factor (TOF) via a multi-window data-resampling scheme. We show this windowing scheme extends the anomaly detection capabilities of the TOF algorithm to resolve noise transients of arbitrary morphology and duration. We demonstrate the efficacy of this pipeline in detecting glitches during LIGO and Virgo's third observing run, and discuss potential applications.