A contrastive self-supervised convolutional autoencoder detects core-collapse supernova gravitational waves with performance comparable to supervised CNNs, better generalization to unseen waveforms, and ~120 kpc sensitive distance under Einstein Telescope noise.
Bayesian reconstruction of gravitational wave bursts using chirplets
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
abstract
The LIGO-Virgo collaboration uses a variety of techniques to detect and characterize gravitational waves. One approach is to use templates - models for the signals derived from Einstein's equations. Another approach is to extract the signals directly from the coherent response of the detectors in LIGO-Virgo network. Both approaches played an important role in the first gravitational wave detections. Here we extend the BayesWave analysis algorithm, which reconstructs gravitational wave signals using a collection of continuous wavelets, to use a generalized wavelet family, known as chirplets, that have time-evolving frequency content. Since generic gravitational wave signals have frequency content that evolves in time, a collection of chirplets provides a more compact representation of the signal, resulting in more accurate waveform reconstructions, especially for low signal-to-noise events, and events that occupy a large time-frequency volume.
citation-role summary
background 1
citation-polarity summary
fields
gr-qc 1years
2026 1verdicts
UNVERDICTED 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
Contrastive self-supervised convolutional autoencoder for core-collapse supernova gravitational-wave detection
A contrastive self-supervised convolutional autoencoder detects core-collapse supernova gravitational waves with performance comparable to supervised CNNs, better generalization to unseen waveforms, and ~120 kpc sensitive distance under Einstein Telescope noise.