Schematic representation of the proposed event classification. The procedure is based on the reconstructed time-frequency map of candidates. ANNs are trained to produce an output number close to 1 for events are classified as belonging to the target distribution, and close to 0 otherwise. Our procedure does not constrain the output value to be limited to [0,1] and overflows and underflows are possible.

Schematic representation of the proposed event classification. The procedure is based on the reconstructed time-frequency map of candidates. ANNs are trained to produce an output number close to 1 for events are classified as belonging to the target distribution, and close to 0 otherwise. Our procedure does not constrain the output value to be limited to [0,1] and overflows and underflows are possible.

Source publication
Article
Full-text available
The quest to observe gravitational waves challenges our ability to discriminate signals from detector noise. This issue is especially relevant for transient gravitational waves searches with a robust eyes wide open approach, the so called all- sky burst searches. Here we show how signal classification methods inspired by broad astrophysical charact...

Citations

... The cWB algorithm has been recently used in combination with machine learning (ML) algorithms for various studies [21][22][23][24]. In this paper, the standard cWB pipeline sensitivity to BBH mergers is enhanced by using the ML method as described in Ref. [25]. ...
Preprint
In this work, we use the coherent WaveBurst (cWB) pipeline enhanced with machine learning (ML) to search for binary black hole (BBH) mergers in the Advanced LIGO-Virgo strain data from the third observing run (O3). We detect, with equivalent or higher significance, all gravitational-wave (GW) events previously reported by the standard cWB search for BBH mergers in the third GW Transient Catalog (GWTC-3). The ML-enhanced cWB search identifies five additional GW candidate events from the catalog that were previously missed by the standard cWB search. Moreover, we identify three marginal candidate events not listed in GWTC-3. For simulated events distributed uniformly in a fiducial volume, we improve the detection efficiency with respect to the standard cWB search by approximately $20\%$ for both stellar-mass and intermediate mass black hole binary mergers, detected with a false-alarm rate less than $1\,\mathrm{yr}^{-1}$. We show the robustness of the ML-enhanced search for detection of generic BBH signals by reporting increased sensitivity to the spin-precessing and eccentric BBH events as compared to the standard cWB search. Furthermore, we compare the improvement of the ML-enhanced cWB search for different detector networks.
... However, there have been ML searches for specific burst source types. In reference [164], the authors employ a neural network algorithm to reduce the impact of glitches on the cWB burst search and increase the significance of the CBC signals which are detected by the pipeline. ...
Article
Full-text available
Machine learning has emerged as a popular and powerful approach for solving problems in astrophysics. We review applications of machine learning techniques for the analysis of ground-based gravitational-wave (GW) detector data. Examples include techniques for improving the sensitivity of Advanced Laser Interferometer GW Observatory and Advanced Virgo GW searches, methods for fast measurements of the astrophysical parameters of GW sources, and algorithms for reduction and characterization of non-astrophysical detector noise. These applications demonstrate how machine learning techniques may be harnessed to enhance the science that is possible with current and future GW detectors.
... More recently it was used for the first time on simulated continuous gravitational wave signals [29] and it was applied to the related long transient signals [30]. Furthermore DNNs have been studied as a follow-up method for CW searches [31,32], as well as for parameter estimation of searches for compact binary merger signals [33,34] and for a multitude of other gravitational-wave-search related applications such as classifying disturbances (glitches) and searches for unmodeled burst signals [35][36][37][38][39][40][41]. ...
Article
Full-text available
The sensitivity of wide-parameter-space searches for continuous gravitational waves is limited by computational cost. Recently it was shown that deep neural networks (DNNs) can perform all-sky searches directly on (single-detector) strain data [C. Dreissigacker , Phys. Rev. D 100, 044009 (2019)], potentially providing a low-computing-cost search method that could lead to a better overall sensitivity. Here we expand on this study in two respects: (i) using (simulated) strain data from two detectors simultaneously, and (ii) training for directed (i.e., single sky-position) searches in addition to all-sky searches. For a data time span of T=105 s, the all-sky two-detector DNN is about 7% less sensitive (in amplitude h0) at low frequency (f=20 Hz), and about 51% less sensitive at high frequency (f=1000 Hz) compared to fully-coherent matched-filtering (using weave). In the directed case the sensitivity gap compared to matched-filtering ranges from about 7%–14% at f=20 Hz to about 37%–49% at f=1500 Hz. Furthermore we assess the DNN’s ability to generalize in signal frequency, spin down and sky-position, and we test its robustness to realistic data conditions, namely gaps in the data and using real LIGO detector noise. We find that the DNN performance is not adversely affected by gaps in the test data or by using a relatively undisturbed band of LIGO detector data instead of Gaussian noise. However, when using a more disturbed LIGO band for the tests, the DNN’s detection performance is substantially degraded due to the increase in false alarms, as expected.
... More recently DNNs have started to draw attention in the field of gravitational-wave searches (i) as a classifier for non-Gaussian detector transients (glitches) [24][25][26][27], (ii) as a search method for unmodeled burst signals [28,29] in time-frequency images produced by coherent WaveBurst [30], and (iii) as a direct detection method for black-hole merger signals in gravitational-wave strain data [31][32][33][34][35][36]. ...
Article
Full-text available
We present a first proof-of-principle study for using deep neural networks (DNNs) as a novel search method for continuous gravitational waves (CWs) from unknown spinning neutron stars. The sensitivity of current wide-parameter-space CW searches is limited by the available computing power, which makes neural networks an interesting alternative to investigate, as they are extremely fast once trained and have recently been shown to rival the sensitivity of matched filtering for black-hole merger signals [D. George and E. A. Huerta, Phys. Rev. D 97, 044039 (2018); H. Gabbard, M. Williams, F. Hayes, and C. Messenger, Phys. Rev. Lett. 120, 141103 (2018)]. We train a convolutional neural network with residual (shortcut) connections and compare its detection power to that of a fully coherent matched-filtering search using the Weave pipeline [K. Wette, S. Walsh, R. Prix, and M. A. Papa, Phys. Rev. D 97, 123016 (2018)]. As test benchmarks we consider two types of all-sky searches over the frequency range from 20 to 1000 Hz: an “easy” search using T=105 s of data, and a “harder” search using T=106 s. The detection probability pdet is measured on a signal population for which matched filtering achieves pdet=90% in Gaussian noise. In the easiest test case (T=105 s at 20 Hz) the DNN achieves pdet∼88%, corresponding to a loss in sensitivity depth of ∼5% versus coherent matched filtering. However, at higher frequencies and for longer observation times the DNN detection power decreases, until pdet∼13% and a loss of ∼66% in sensitivity depth in the hardest case (T=106 s at 1000 Hz). We study the DNN generalization ability by testing on signals of different frequencies, spindowns and signal strengths than they were trained on. We observe excellent generalization: only five networks, each trained at a different frequency, would be able to cover the whole frequency range of the search.
... More recently DNNs have started to draw attention in the field of gravitational-wave searches: (i) as a classifier for non-Gaussian detector transients (glitches) [26][27][28][29], (ii) as a search method for unmodelled burst signals [30,31] in time-frequency images produced by coherent WaveBurst [32], and (iii) as a direct detection method for black-hole merger signals in gravitational-wave strain data [1,2,[33][34][35][36]. ...
Preprint
Full-text available
We present a first proof-of-principle study for using deep neural networks (DNNs) as a novel search method for continuous gravitational waves (CWs) from unknown spinning neutron stars. The sensitivity of current wide-parameter-space CW searches is limited by the available computing power, which makes neural networks an interesting alternative to investigate, as they are extremely fast once trained and have recently been shown to rival the sensitivity of matched filtering for black-hole merger signals. We train a convolutional neural network with residual (short-cut) connections and compare its detection power to that of a fully-coherent matched-filtering search using the WEAVE pipeline. As test benchmarks we consider two types of all-sky searches over the frequency range from $20\,\mathrm{Hz}$ to $1000\,\mathrm{Hz}$: an `easy' search using $T=10^5\,\mathrm{s}$ of data, and a `harder' search using $T=10^6\,\mathrm{s}$. Detection probability $p_\mathrm{det}$ is measured on a signal population for which matched filtering achieves $p_\mathrm{det}=90\%$ in Gaussian noise. In the easiest test case ($T=10^5\,\mathrm{s}$ at $20\,\mathrm{Hz}$) the DNN achieves $p_\mathrm{det}\sim88\%$, corresponding to a loss in sensitivity depth of $\sim5\%$ versus coherent matched filtering. However, at higher-frequencies and longer observation time the DNN detection power decreases, until $p_\mathrm{det}\sim13\%$ and a loss of $\sim 66\%$ in sensitivity depth in the hardest case ($T=10^6\,\mathrm{s}$ at $1000\,\mathrm{Hz}$). We study the DNN generalization ability by testing on signals of different frequencies, spindowns and signal strengths than they were trained on. We observe excellent generalization: only five networks, each trained at a different frequency, would be able to cover the whole frequency range of the search.
Article
Coherent WaveBurst is a generic, multidetector gravitational wave burst search based on the excess power approach. The coherent WaveBurst algorithm currently employed in the all-sky short-duration gravitational wave burst search uses a conditional approach on selected attributes in the multidimensional event attribute space to distinguish between noisy events from that of astrophysical origin. We have been developing a supervised machine learning approach based on the Gaussian mixture modeling to model the attribute space for signals as well as noise events to enhance the probability of burst detection [Gayathri et al.Phys. Rev. D 102, 104023 (2020)]. We further extend the Gaussian mixture model approach to the all-sky short-duration coherent WaveBurst search as a postprocessing step on events from the first half of the third observing run (O3a). We show an improvement in sensitivity to generic gravitational wave burst signal morphologies as well as the astrophysical source such as core-collapse supernova models due to the application of our Gaussian mixture model approach to coherent WaveBurst triggers. The Gaussian mixture model method recovers the gravitational wave signals from massive compact binary coalescences identified by coherent WaveBurst targeted for binary black holes in GWTC-2, with better significance than the all-sky coherent WaveBurst search. No additional significant gravitational wave bursts are observed.
Article
The coherent WaveBurst (cWB) search algorithm identifies generic gravitational wave (GW) signals in the LIGO-Virgo strain data. We propose a machine learning (ML) method to optimize the pipeline sensitivity to the special class of GW signals known as binary black hole (BBH) mergers. Here, we test the ML-enhanced cWB search on strain data from the first and second observing runs of Advanced LIGO and successfully recover all BBH events previously reported by cWB, with higher significance. For simulated events found with a false alarm rate less than 1 yr−1, we demonstrate the improvement in the detection efficiency of 26% for stellar-mass BBH mergers and 16% for intermediate mass black hole binary mergers. To demonstrate the robustness of the ML-enhanced search for the detection of generic BBH signals, we show that it has the increased sensitivity to the spin precessing or eccentric BBH events, even when trained on simulated quasicircular BBH events with aligned spins.
Article
Full-text available
This work investigates the problem of detecting gravitational wave (GW) events based on simulated damped sinusoid signals contaminated with white Gaussian noise. It is treated as a classification problem with one class for the interesting events. The proposed scheme consists of the following two successive steps: decomposing the data using a wavelet packet, representing the GW signal and noise using the derived decomposition coeficients; and determining the existence of any GW event using a convolutional neural network (CNN) with a logistic regression output layer. The characteristic of this work is its comprehensive investigations on CNN structure, detection window width, data resolution, wavelet packet decomposition and detection window overlap scheme. Extensive simulation experiments show excellent performances for reliable detection of signals with a range of GW model parameters and signal-to-noise ratios. While we use a simple waveform model in this study, we expect the method to be particularly valuable when the potential GW shapes are too complex to be characterized with a template bank.
Article
The LIGO observatories detect gravitational waves through monitoring changes in the detectors’ length down to below 10−19 m/Hz variations—a small fraction of the size of the atoms that make up the detector. To achieve this sensitivity, the detector and its environment need to be closely monitored. Beyond the gravitational-wave data stream, LIGO continuously records hundreds of thousands of channels of environmental and instrumental data in order to monitor for possibly minuscule variations that contribute to the detector noise. A particularly challenging issue is the appearance in the gravitational wave signal of brief, loud noise artifacts called “glitches,” which are environmental or instrumental in origin but can mimic true gravitational waves and therefore hinder sensitivity. Currently, they are primarily identified by analysis of the gravitational-wave data stream, and auxiliary data channels often provide corroborating evidence. Here we present a machine learning approach that can identify glitches by considering all environmental and detector data channels, a task that has not previously been pursued due to its scale and the number of degrees of freedom within gravitational-wave detectors. The presented method is capable of reducing the gravitational-wave detector network’s false alarm rate and improving the LIGO instruments, consequently enhancing detection confidence.
Article
In this article, I introduce ideas and techniques to extract information about the equation of state of matter at very high densities from gravitational waves emitted before, during and after the merger of binary neutron stars. I also review current work and results on the actual use of the first gravitational-wave observation of a neutron-star merger to set constraints on properties of such equation of state. In passing, I also touch on the possibility that what we observe in gravitational waves are not neutron stars, but something more exotic. In order to make this article more accessible, I also review the dynamics and gravitational-wave emission of neutron-star mergers in general, with focus on numerical simulations and on which representations of the equation of state are used for studies on binary systems.