arXiv subjects

Chris Messenger

Publications and source records attributed to Chris Messenger.

At least 37 records · Page 2Linked to original sources

arXiv2020

A robust machine learning algorithm to search for continuous gravitational waves

Many continuous gravitational wave searches are affected by instrumental spectral lines that could be confused with a continuous astrophysical signal. Several techniques have been developed to limit the effect of these lines by penalising signals that appear in only a single detector. We have developed a general method, using a convolutional neural network, to reduce the impact of instrumental artefacts on searches that use the SOAP algorithm. The method can identify features in corresponding frequency bands of each detector and classify these bands as containing a signal, an instrumental line, or noise. We tested the method against four different data-sets: Gaussian noise with time gaps, data from the final run of Initial LIGO (S6) with signals added, the reference S6 mock data challenge data set and signals injected into data from the second advanced LIGO observing run (O2). Using the S6 mock data challenge data set and at a 1% false alarm probability we showed that at 95% efficiency a fully-automated SOAP search has a sensitivity corresponding to a coherent signal-to-noise ratio of 110, equivalent to a sensitivity depth of 10 Hz$^{-1/2}$, making this automated search competitive with other searches requiring significantly more computing resources and human intervention.

astro-ph.IM↗

arXiv2020

Enhancing the sensitivity of transient gravitational wave searches with Gaussian Mixture Models

Identifying the presence of a gravitational wave transient buried in non-stationary, non-Gaussian noise which can often contain spurious noise transients (glitches) is a very challenging task. For a given data set, transient gravitational wave searches produce a corresponding list of triggers that indicate the possible presence of a gravitational wave signal. These triggers are often the result of glitches mimicking gravitational wave signal characteristics. To distinguish glitches from genuine gravitational wave signals, search algorithms estimate a range of trigger attributes, with thresholds applied to these trigger properties to separate signal from noise. Here, we present the use of Gaussian mixture models, a supervised machine learning approach, as a means of modelling the multi-dimensional trigger attribute space. We demonstrate this approach by applying it to triggers from the coherent Waveburst search for generic bursts in LIGO O1 data. By building Gaussian mixture models for the signal and background noise attribute spaces, we show that we can significantly improve the sensitivity of the coherent Waveburst search and strongly suppress the impact of glitches and background noise, without the use of multiple search bins as employed by the original O1 search. We show that the detection probability is enhanced by a factor of 10, leading enhanced statistical significance for gravitational wave signals such as GW150914.

gr-qc↗

arXiv2019

Detection and Classification of Supernova Gravitational Waves Signals: A Deep Learning Approach

We demonstrate the application of a convolutional neural network to the gravitational wave signals from core collapse supernovae. Using simulated time series of gravitational wave detectors, we show that based on the explosion mechanisms, a convolutional neural network can be used to detect and classify the gravitational wave signals buried in noise. For the waveforms used in the training of the convolutional neural network, our results suggest that a network of advanced LIGO, advanced VIRGO and KAGRA, or a network of LIGO A+, advanced VIRGO and KAGRA is likely to detect a magnetorotational core collapse supernovae within the Large and Small Magellanic Clouds, or a Galactic event if the explosion mechanism is the neutrino-driven mechanism. By testing the convolutional neural network with waveforms not used for training, we show that the true alarm probabilities are 52% and 83% at 60 kpc for waveforms R3E1AC and R4E1FC L. For waveforms s20 and SFHx at 10 kpc, the true alarm probabilities are 70% and 93% respectively. All at false alarm probability equal to 10%.

astro-ph.HE↗

arXiv2019

Deep-Learning continuous gravitational waves

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.

gr-qc↗

arXiv2019

SOAP: A generalised application of the Viterbi algorithm to searches for continuous gravitational-wave signals

All-sky and wide parameter space searches for continuous gravitational waves are generally template-matching schemes which test a bank of signal waveforms against data from a gravitational wave detector. Such searches can offer optimal sensitivity for a given computing cost and signal model, but are highly-tuned to specific signal types and are computationally expensive, even for semi-coherent searches. We have developed a search method based on the well-known Viterbi algorithm which is model-agnostic and has a computational cost several orders of magnitude lower than template methods, with a modest reduction in sensitivity. In particular, this method can search for signals which have an unknown frequency evolution. We test the algorithm on three simulated and real data sets: gapless Gaussian noise, Gaussian noise with gaps and real data from the final run of initial LIGO (S6). We show that at 95% efficiency, with a 1% false alarm rate, the algorithm has a depth sensitivity of $\sim 33$, $10$ and $13$ ,Hz$^{-1/2}$ with corresponding SNRs of $\sim 60$, $72$ and $74$ in these datasets. we discuss the use of this algorithm for detecting a wide range of quasi-monochromatic gravitational wave signals and instrumental lines.

astro-ph.IM↗

arXiv2018

Binary Neutron Star Mergers and Third Generation Detectors: Localization and Early Warning

For third generation gravitational wave detectors, such as the Einstein Telescope, gravitational wave signals from binary neutron stars can last up to a few days before the neutron stars merge. To estimate the measurement uncertainties of key signal parameters, we develop a Fisher matrix approach which accounts for effects on such long duration signals of the time-dependent detector response and the earths rotation. We use this approach to characterize the sky localization uncertainty for gravitational waves from binary neutron stars at 40, 200, 400, 800 and 1600Mpc, for the Einstein Telescope and Cosmic Explorer individually and operating as a network. We find that the Einstein Telescope alone can localize the majority of detectable binary neutron stars at a distance of $\leq200$Mpc to within $100\text{deg}^2$ with 90% confidence. A network consisting of the Einstein Telescope and Cosmic Explorer can enhance the sky localization performance significantly - with the 90% credible region of $\mathcal{O}(1) \text{deg}^2$ for most sources at $\leq200$Mpc and $\leq100\text{deg}^2$ for most sources at $\leq1600$Mpc. We also investigate the prospects for third generation detectors identifying the presence of a signal prior to merger. To do this, we require a signal to have a network signal-to-noise ratio of $\geq12$ and $\geq5.5$ for at least two interferometers, and to have a 90% credible region for the sky localization that is no larger than $100 \text{deg}^2$. We find that the Einstein Telescope can send out such "early-warning" detection alerts 1 - 20 hours before merger for 100% of detectable binary neutron stars at 40Mpc and for $\sim58\%$ of sources at 200Mpc. For sources at a distance of 400Mpc, a network of the Einstein telescope and Cosmic Explorer can produce detection alerts up to $\sim 3$ hours prior to merger for 98% of detectable binary neutron stars.

astro-ph.HE↗

arXiv2018

Strategies for the Follow-up of Gravitational Wave Transients with the Cherenkov Telescope Array

The observation of the electromagnetic counterpart of gravitational-wave (GW) transient GW170817 demonstrated the potential in extracting astrophysical information from multimessenger discoveries. The forthcoming deployment of the first telescopes of the Cherenkov Telescope Array (CTA) observatory will coincide with Advanced LIGO/Virgo's next observing run, O3, enabling the monitoring of gamma-ray emission at E > 20 GeV, and thus particle acceleration, from GW sources. CTA will not be greatly limited by the precision of GW localization as it will be be capable of rapidly covering the GW error region with sufficient sensitivity. We examine the current status of GW searches and their follow-up effort, as well as the status of CTA, in order to identify some of the general strategies that will enhance CTA's contribution to multimessenger discoveries.

astro-ph.HE↗

arXiv2017

Matching matched filtering with deep networks in gravitational-wave astronomy

We report on the construction of a deep convolutional neural network that can reproduce the sensitivity of a matched-filtering search for binary black hole gravitational-wave signals. The standard method for the detection of well modeled transient gravitational-wave signals is matched filtering. However, the computational cost of such searches in low latency will grow dramatically as the low frequency sensitivity of gravitational-wave detectors improves. Convolutional neural networks provide a highly computationally efficient method for signal identification in which the majority of calculations are performed prior to data taking during a training process. We use only whitened time series of measured gravitational-wave strain as an input, and we train and test on simulated binary black hole signals in synthetic Gaussian noise representative of Advanced LIGO sensitivity. We show that our network can classify signal from noise with a performance that emulates that of match filtering applied to the same datasets when considering the sensitivity defined by Reciever-Operator characteristics.

astro-ph.IM↗

arXiv2017

Accretion-induced spin-wandering effects on the neutron star in Scorpius X-1: Implications for continuous gravitational wave searches

The LIGO's discovery of binary black hole mergers has opened up a new era of transient gravitational wave astronomy. The potential detection of gravitational radiation from another class of astronomical objects, rapidly spinning non-axisymmetric neutron stars, would constitute a new area of gravitational wave astronomy. Scorpius X-1 (Sco X-1) is one of the most promising sources of continuous gravitational radiation to be detected with present-generation ground-based gravitational wave detectors, such as Advanced LIGO and Advanced Virgo. As the sensitivity of these detectors improve in the coming years, so will power of the search algorithms being used to find gravitational wave signals. Those searches will still require integation over nearly year long observational spans to detect the incredibly weak signals from rotating neutron stars. For low mass X-ray binaries such as Sco X-1 this difficult task is compounded by neutron star "spin wandering" caused by stochastic accretion fluctuations. In this paper, we analyze X-ray data from the RXTE satellite to infer the fluctuating torque on the neutron star in Sco X-1. We then perform a large-scale simulation to quantify the statistical properties of spin-wandering effects on the gravitational wave signal frequency and phase evolution. We find that there are a broad range of expected maximum levels of frequency wandering corresponding to maximum drifts of between 0.3-50 μHz/sec over a year at 99% confidence. These results can be cast in terms of the maximum allowed length of a coherent signal model neglecting spin-wandering effects as ranging between 5-80 days. This study is designed to guide the development and evaluation of Sco X-1 search algorithms.

gr-qc↗

arXiv2017

Cosmological inference using only gravitational wave observations of binary neutron stars

[Abridged] This study presents the first Bayesian investigation of the accuracy with which the cosmological parameters can be measured using information coming \emph{only} from the gravitational wave observations of binary neutron star systems by Einstein Telescope. We find, by direct simulation of $10^3$ detections of binary neutron stars, that, within our simplifying assumptions, $H_0,Ω_m,Ω_Λ,w_0$ and $w_1$ can be measured at the $95\%$ level with an accuracy of $\sim 8\%,65\%,39\%,80\%$ and $90\%$, respectively. We also find, by extrapolation, that a measurement accuracy comparable with current measurements by Planck is possible if the number of gravitational wave events observed is $O(10^{6-7})$. We conclude that, while not competitive with electro-magnetic missions in terms of significant digits, gravitational wave alone are capable of providing a complementary determination of the dynamics of the Universe.

gr-qc↗

arXiv2016

Maximising the detection probability of kilonovae associated with gravitational wave observations

Estimates of the source sky location for gravitational wave signals are likely span areas ranging up to hundreds of square degrees or more, making it very challenging for most telescopes to search for counterpart signals in the electromagnetic spectrum. To boost the chance of successfully observing such counterparts, we have developed an algorithm which optimizes the number of observing fields and their corresponding time allocations by maximizing the detection probability. As a proof-of-concept demonstration, we optimize follow-up observations targeting kilonovae using telescopes including CTIO-Dark Energy Camera, Subaru-HyperSuprimeCam, Pan-STARRS and Palomar Transient Factory. We consider three simulated gravitational wave events with 90% credible error regions spanning areas from ~30 deg^2 to ~300 deg^2. Assuming a source at 200 Mpc, we demonstrate that to obtain a maximum detection probability, there is an optimized number of fields for any particular event that a telescope should observe. To inform future telescope design studies, we present the maximum detection probability and corresponding number of observing fields for a combination of limiting magnitudes and fields-of-view over a range of parameters. We show that for large gravitational wave error regions, telescope sensitivity rather than field-of-view, is the dominating factor in maximizing the detection probability.

astro-ph.HE↗

arXiv2016

Systematic errors in estimation of gravitational-wave candidate significance

We investigate the issue in determining the significance of candidate transient gravitational-wave events in a ground-based interferometer network. Given the presence of non-Gaussian noise artefacts in real data, the noise background must be estimated empirically from the data itself. However, the data also potentially contains signals, thus the background estimate may be overstated due to contributions from signals. It has been proposed to mitigate possible bias by removing single-detector data samples that pass a multi-detector consistency test from the background estimates. We conduct a high-statistics Mock Data Challenge to evaluate the effects of removing such samples, modelling a range of scenarios with plausible detector noise distributions and with a range of plausible foreground astrophysical signal rates. We consider the two different modes: one in which coincident samples are removed, and one in which all samples are retained and used. Three algorithms were operated in both modes, show good consistency with each other; however, discrepancies arise between the results obtained under the "coincidence removal" and "all samples" modes, for false alarm probabilities below a certain value. In most scenarios the median of the false alarm probability (FAP) estimator under the "all samples" mode is consistent with the exact FAP. On the other hand the "coincidence removal" mode is found to be unbiased for the mean of the estimated FAP. While the numerical values at which discrepancies become apparent are specific to the details of our experiment, we believe that the qualitative differences in the behaviour of the median and mean of the FAP estimator have more general validity. On the basis of our study we suggest that the FAP of candidates for the first detection of gravitational waves should be estimated without removing single-detector samples that form coincidences.

astro-ph.IM↗

arXiv2014

A semi-coherent search for weak pulsations in Aql X-1

Non pulsating neutron stars in low mass X-ray binaries largely outnumber those that show pulsations. The lack of detectable pulses represents a big open problem for two important reasons. The first is that the structure of the accretion flow in the region closest to the neutron star is not well understood and it is therefore unclear what is the mechanism that prevents the pulse formation. The second is that the detection of pulsations would immediately reveal the spin of the neutron star. Aql X-1 is a special source among low mass X-ray binaries because it has showed the unique property of pulsating for only ~150 seconds out of a total observing time of more than 1.5 million seconds. However, the existing upper limits on the pulsed fraction leave open two alternatives. Either Aql X-1 has very weak pulses which have been undetected, or it has genuinely pulsed only for a tiny amount of the observed time. Understanding which of the two scenarios is the correct one is fundamental to increase our knowledge about the pulse formation process and understand the chances we have to detect weak pulses in other LMXBs. In this paper we perform a semi-coherent search on the entire X-ray data available for Aql X-1. We find no evidence for (new) weak pulsations with the most stringent upper limits being of the order of 0.3% in the 7-25 keV energy band.

astro-ph.HE↗

arXiv2014

Data Analysis Methods for Testing Alternative Theories of Gravity with LISA Pathfinder

In this paper we present a data analysis approach applicable to the potential saddle-point fly-by mission extension of LISA Pathfinder (LPF). At the peak of its sensitivity, LPF will sample the gravitational field in our Solar System with a precision of several $\text{fm/s}^2/\sqrt{\text{Hz}}$ at frequencies around $1\,\text{mHz}$. Such an accurate accelerometer will allow us to test alternative theories of gravity that predict deviations from Newtonian dynamics in the non-relativistic limit. As an example, we consider the case of the Tensor-Vector-Scalar theory of gravity and calculate, within the non-relativistic limit of this theory, the signals that anomalous tidal stresses generate in LPF. We study the parameter space of these signals and divide it into two subgroups, one related to the mission parameters and the other to the theory parameters that are determined by the gravity model. We investigate how the mission parameters affect the signal detectability concluding that these parameters can be determined with the sufficient precision from the navigation of the spacecraft and fixed during our analysis. Further, we apply Bayesian parameter estimation and determine the accuracy to which the gravity theory parameters may be inferred. We evaluate the portion of parameter space that may be eliminated in case of no signal detection and estimate the detectability of signals as a function of parameter space location. We also perform a first investigation of non-Gaussian "noise-glitches" that may occur in the data. The analysis we develop is universal and may be applied to anomalous tidal stress induced signals predicted by any theory of gravity.

gr-qc↗

arXiv2013

Host redshifts from gravitational-wave observations of binary neutron star mergers

Inspiralling compact binaries as standard sirens will soon become an invaluable tool for cosmology when advanced interferometric gravitational-wave detectors begin their observations in the coming years. However, a degeneracy in the information carried by gravitational waves between the total rest-frame mass $M$ and the redshift $z$ of the source implies that neither can be directly extracted from the signal, but only the combination $M(1+z)$, the redshifted mass. Recent work has shown that for binary neutron star systems, a tidal correction to the gravitational-wave phase in the late-inspiral signal that depends on the rest-frame source mass could be used to break the mass-redshift degeneracy. We propose here to use the signature encoded in the post-merger signal to deduce the redshift to the source. This will allow an accurate extraction of the intrinsic rest-frame mass of the source, in turn permitting the determination of source redshift and luminosity distance solely from gravitational-wave observations. This will herald a new era in precision cosmography and astrophysics. Using numerical simulations of binary neutron star mergers of very slightly different mass, we model gravitational-wave signals at different redshifts and use Bayesian parameter estimation to determine the accuracy with which the redshift can be extracted for a source of known mass. We find that the Einstein Telescope can determine the source redshift to $sim 10$--$20%$ at redshifts of $z<0.04$.

gr-qc↗

arXiv2011

Measuring a cosmological distance-redshift relationship using only gravitational wave observations of binary neutron star coalescences

Detection of gravitational waves from the inspiral phase of binary neutron star coalescence will allow us to measure the effects of the tidal coupling in such systems. These effects will be measurable using 3rd generation gravitational wave detectors, e.g. the Einstein Telescope, which will be capable of detecting inspiralling binary neutron star systems out to redshift z=4. Tidal effects provide additional contributions to the phase evolution of the gravitational wave signal that break a degeneracy between the system's mass parameters and redshift and thereby allow the simultaneous measurement of both the effective distance and the redshift for individual sources. Using the population of O(10^3-10^7) detectable binary neutron star systems predicted for the Einstein Telescope the luminosity distance--redshift relation can be probed independently of the cosmological distance ladder and independently of electromagnetic observations. We present the results of a Fisher information analysis applied to waveforms assuming a subset of possible neutron star equations of state. We conclude that for our range of representative neutron star equations of state the redshift of such systems can be determined to an accuracy of 8-40% for z<1 and 9-65% for 1<z<4.

gr-qc↗

arXiv2011

Bayesian versus frequentist upper limits

While gravitational waves have not yet been measured directly, data analysis from detection experiments commonly includes an upper limit statement. Such upper limits may be derived via a frequentist or Bayesian approach; the theoretical implications are very different, and on the technical side, one notable difference is that one case requires maximization of the likelihood function over parameter space, while the other requires integration. Using a simple example (detection of a sinusoidal signal in white Gaussian noise), we investigate the differences in performance and interpretation, and the effect of the "trials factor", or "look-elsewhere effect".

physics.data-an↗

arXiv2011

A semi-coherent search strategy for known continuous wave sources in binary systems

We present a method for detection of weak continuous signals from sources in binary systems via the incoherent combination of many "short" coherently-analyzed segments. The main focus of the work is on the construction of a metric on the parameter space for such signals for use in matched-filter based searches. The metric is defined using a maximum likelihood detection statistic applied to a binary orbit phase model including eccentricity. We find that this metric can be accurately approximated by its diagonal form in the regime where the segment length is << the orbital period. Hence correlations between parameters are effectively removed by the combination of many independent observation. We find that the ability to distinguish signal parameters is independent of the total semi-coherent observation span (for the semi-coherent span >> the segment length) for all but the orbital angular frequency. Increased template density for this parameter scales linearly with the observation span. We also present two example search schemes. The first uses a re parameterized phase model upon which we compute the metric on individual short coherently analyzed segments. The second assumes long >> the orbital period segment lengths from which we again compute the coherent metric and find it to be approximately diagonal. In this latter case we also show that the semi-coherent metric is equal to the coherent metric.

gr-qc↗