Speaker
Mr
Julian Ding
(University of British Columbia)
Description
The data from current gravitational wave detectors, including Advanced LIGO and Advanced Virgo, contains a high rate of “glitches”: transient noise that can obscure true detections of gravitational wave events. I propose a novel method for identifying and characterizing these noise events by leveraging the Temporal Outlier Factor (TOF): a technique that uses higher-dimensional embedding to recreate the dynamical phase space of a time series, then correlates spatial and temporal clustering to find unique anomalies in the data. I present preliminary results demonstrating TOF’s efficacy in detecting known glitch classes in Advanced LIGO detector data, then discuss potential applications for detecting unknown glitch types.
Author
Mr
Julian Ding
(University of British Columbia)
Co-authors
Dr
Jess McIver
(University of British Columbia)
Dr
Raymond Ng
(University of British Columbia)
