GWAK: Gravitational-Wave Anomalous Knowledge with Recurrent Autoencoders

• DOI: 10.48550/arxiv.2309.11537
• Publication Date: 2024-04-25
• Authors: Ryan Raikman; Eric A Moreno; Ekaterina Govorkova; Ethan J Marx; Alec Gunny; William Benoit; Deep Chatterjee; Rafia Omer; Muhammed Saleem; Dylan S Rankin; Michael W Coughlin; Philip C Harris; Erik Katsavounidis
• Keywords: semi-supervised learning; Computer engineering. Computer hardware; FOS: Physical sciences; gravitational-wave physics; QA75.5-76.95; General Relativity and Quantum Cosmology (gr-qc); 01 natural sciences; anomaly detection; General Relativity and Quantum Cosmology; TK7885-7895; machine learning; autoencoders; Electronic computers. Computer science; 0103 physical sciences; Astrophysics - Instrumentation and Methods for Astrophysics; Instrumentation and Methods for Astrophysics (astro-ph.IM)

Abstract Matched-filtering detection techniques for gravitational-wave (GW) signals in ground-based interferometers rely on having well-modeled templates of the GW emission. Such techniques have been traditionally used in searches for compact binary coalescences (CBCs), and have been employed in all known GW detections so far. However, interesting science cases aside from compact mergers do not yet have accurate enough modeling to make matched filtering possible, including core-collapse supernovae and sources where stochasticity may be involved. Therefore the development of techniques to identify sources of these types is of significant interest. In this paper, we present a method of anomaly detection based on deep recurrent autoencoders to enhance the search region to unmodeled transients. We use a semi-supervised strategy that we name ‘Gravitational Wave Anomalous Knowledge’ (GWAK). While the semi-supervised approach to this problem entails a potential reduction in accuracy compared to fully supervised methods, it offers a generalizability advantage by enhancing the reach of experimental sensitivity beyond the constraints of pre-defined signal templates. We construct a low-dimensional embedded space using the GWAK method, capturing the physical signatures of distinct signals on each axis of the space. By introducing signal priors that capture some of the salient features of GW signals, we allow for the recovery of sensitivity even when an unmodeled anomaly is encountered. We show that regions of the GWAK space can identify CBCs, detector glitches and also a variety of unmodeled astrophysical sources.