Simulating Transient Noise Bursts in LIGO with Generative Adversarial Networks

Abstract

The noise of gravitational-wave (GW) interferometers limits their sensitivity and impacts the data quality, hindering the detection of GW signals from astrophysical sources. For transient searches, the most problematic are transient noise artifacts, known as glitches, that happen at a rate around 1 min −1 , and can mimic GW signals. Because of this, there is a need for better modeling and inclusion of glitches in large-scale studies, such as stress testing the pipelines. In this proof-of concept work we employ Generative Adversarial Networks (GAN), a state-of-the-art Deep Learning algorithm inspired by Game Theory, to learn the underlying distribution of blip glitches and to generate artificial populations. We reconstruct the glitch in the time-domain, providing a smooth input that the GAN can learn. With this methodology, we can create distributions of ∼ 10 3 glitches from Hanford and Livingston detectors in less than one second. Furthermore, we employ several metrics to measure the performance of our methodology and the quality of its generations. This investigation will be extended in the future to different glitch classes with the final goal of creating an open-source interface for mock data generation.

Active learning BSM parameter spaces

“…The advantage of our approach, however, is its relative simplicity. With active learning, we only have one model's hyperparamters to tune -finetuning the two competing networks of a GAN is notoriously difficult [38,39] -and we need to train less, meaning that we potentially use less compute. And despite the simplicity, we find compelling accuracies even in higher-dimensional models, as can be seen in sections 5 and 6.…”

Generating transient noise artifacts in gravitational-wave detector data with generative adversarial networks

“…In Refs. [22,23], generative adversarial networks (GANs) [24] are used to produce fake time-series waveforms for the most common type of detector glitch, known as blips. GANs are a type of neural network that can be used for creating new, synthetic data that mimic their input.…”

Convolutional neural network for gravitational-wave early alert: Going down in frequency

“…The main advantage of ML techniques is their rapidity because most of the computations are made during the training stage. A widely used ML method for pattern recognition is based on convolutional neural networks (CNNs) [34], in the context of GW it has been applied to different tasks such as CBC identification [35][36][37][38][39], burst detection [40][41][42][43], sky localization [44][45][46], glitch classification [47,48] and synthetic data generation [49,50]. See [51] for a review on this topic.…”