arXiv subjects

Collin D. Capano

Publications and source records attributed to Collin D. Capano.

33 records · Page 2Linked to original sources

arXiv2021

Genetic-algorithm-optimized neural networks for gravitational wave classification

Gravitational-wave detection strategies are based on a signal analysis technique known as matched filtering. Despite the success of matched filtering, due to its computational cost, there has been recent interest in developing deep convolutional neural networks (CNNs) for signal detection. Designing these networks remains a challenge as most procedures adopt a trial and error strategy to set the hyperparameter values. We propose a new method for hyperparameter optimization based on genetic algorithms (GAs). We compare six different GA variants and explore different choices for the GA-optimized fitness score. We show that the GA can discover high-quality architectures when the initial hyperparameter seed values are far from a good solution as well as refining already good networks. For example, when starting from the architecture proposed by George and Huerta, the network optimized over the 20-dimensional hyperparameter space has 78% fewer trainable parameters while obtaining an 11% increase in accuracy for our test problem. Using genetic algorithm optimization to refine an existing network should be especially useful if the problem context (e.g. statistical properties of the noise, signal model, etc) changes and one needs to rebuild a network. In all of our experiments, we find the GA discovers significantly less complicated networks as compared to the seed network, suggesting it can be used to prune wasteful network structures. While we have restricted our attention to CNN classifiers, our GA hyperparameter optimization strategy can be applied within other machine learning settings.

gr-qc↗

arXiv2020

Binary black hole spectroscopy: a no-hair test of GW190814 and GW190412

Gravitational waves provide a window to probe general relativity (GR) under extreme conditions. The recent observations of GW190412 and GW190814 are unique high-mass-ratio mergers that enable the observation of gravitational-wave harmonics beyond the dominant $(\ell, m) = (2, 2)$ mode. Using these events, we search for physics beyond GR by allowing the source parameters measured from the sub-dominant harmonics to deviate from that of the dominant mode. All results are consistent with GR. We constrain the chirp mass as measured by the $(\ell, m) = (3, 3)$ mode to be within $0_{-3}^{+5}\%$ of the dominant mode when we allow both the masses and spins of the sub-dominant modes to deviate. If we allow only the mass parameters to deviate, we constrain the chirp mass of the $(3, 3)$ mode to be within $\pm1\%$ of the expected value from GR.

gr-qc↗

arXiv2020

GW190521 may be an intermediate mass ratio inspiral

GW190521 is the first confident observation of a binary black hole merger with total mass $M > 100\,\mathrm{M}_{\odot}$. Given the lack of observational constraints at these masses, we analyze GW190521 considering two different priors for the binary's masses: uniform in mass ratio and source-frame total mass, and uniform in source-frame component masses. For the uniform in mass-ratio prior, we find that the component masses are $m_1^{\mathrm{src}} = 168_{-61}^{+15}\,\mathrm{M}_{\odot}$ and $m_2^{\mathrm{src}} = 16_{-3}^{+33}\,\mathrm{M}_{\odot}$. The uniform in component-mass prior yields a bimodal posterior distribution. There is a low-mass-ratio mode ($q<4$) with $m_1^{\mathrm{src}} = 100_{-18}^{+17}\,\mathrm{M}_{\odot}$ and $m_2^{\mathrm{src}} = 57_{-16}^{+17}\,\mathrm{M}_{\odot}$ and a high-mass-ratio mode ($q\geq4$) with $m_1^{\mathrm{src}} = 166_{-35}^{+16}\,\mathrm{M}_{\odot}$ and $m_2^{\mathrm{src}} = 16_{-3}^{+14}\,\mathrm{M}_{\odot}$. Although the two modes have nearly equal posterior probability, the maximum-likelihood parameters are in the high-mass ratio mode, with $m_1^{\rm src} = 171\,M_{\odot}$ and $m_2^{\rm src} = 16\,M_{\odot}$, and a signal-to-noise ratio of $16$. These results are consistent with the proposed "mass gap" produced by pair-instability in supernova. Our results differ from those published in Abbott et al. (2020b). We find that a combination of the prior used and the constraints applied may have prevented that analysis from sampling the high-mass-ratio mode. An accretion flare in AGN J124942.3+344929 was observed in possible coincidence with GW190521 by the Zwicky Transient Facility (ZTF). We report parameters assuming a common origin; however, the spatial agreement of GW190521 and the EM flare alone does not provide convincing evidence for the association ($\ln\mathcal{B} \gtrsim -4$).

astro-ph.HE↗

arXiv2020

Stringent constraints on neutron-star radii from multimessenger observations and nuclear theory

The properties of neutron stars are determined by the nature of the matter that they contain. These properties can be constrained by measurements of the star's size. We obtain stringent constraints on neutron-star radii by combining multimessenger observations of the binary neutron-star merger GW170817 with nuclear theory that best accounts for density-dependent uncertainties in the equation of state. We construct equations of state constrained by chiral effective field theory and marginalize over these using the gravitational-wave observations. Combining this with the electromagnetic observations of the merger remnant that imply the presence of a short-lived hyper-massive neutron star, we find that the radius of a $1.4\,\rm{M}_\odot$ neutron star is $R_{1.4\,\mathrm{M}_\odot} = 11.0^{+0.9}_{-0.6}~{\rm km}$ (90% credible interval). Using this constraint, we show that neutron stars are unlikely to be disrupted in neutron-star black-hole mergers; subsequently, such events will not produce observable electromagnetic emission.

astro-ph.HE↗

arXiv2020

2-OGC: Open Gravitational-wave Catalog of binary mergers from analysis of public Advanced LIGO and Virgo data

We present the second Open Gravitational-wave Catalog (2-OGC) of compact-binary coalescences, obtained from the complete set of public data from Advanced LIGO's first and second observing runs. For the first time we also search public data from the Virgo observatory. The sensitivity of our search benefits from updated methods of ranking candidate events including the effects of non-stationary detector noise and varying network sensitivity; in a separate targeted binary black hole merger search we also impose a prior distribution of binary component masses. We identify a population of 14 binary black hole merger events with probability of astrophysical origin $> 0.5$ as well as the binary neutron star merger GW170817. We confirm the previously reported events GW170121, GW170304, and GW170727 and also report GW151205, a new marginal binary black hole merger with a primary mass of $67^{+28}_{-17}\,\mathrm{M}_{\odot}$ that may have formed through hierarchical merger. We find no additional significant binary neutron star merger or neutron star--black hole merger events. To enable deeper follow-up as our understanding of the underlying populations evolves, we make available our comprehensive catalog of events, including the sub-threshold population of candidates and posterior samples from parameter inference of the 30 most significant binary black hole candidates.

astro-ph.HE↗

arXiv2019

The next decade of black hole spectroscopy

Gravitational wave observations of the ringdown of the remnant black hole in a binary black hole coalescence provide a unique opportunity of confronting the black hole no-hair theorem in general relativity with observational data. The most robust tests are possible if multiple ringdown modes can be observed. In this paper, using state-of-the-art Bayesian inference methods and the most up-to-date knowledge of binary black hole population parameters and ringdown mode amplitudes, we evaluate the prospects for black hole spectroscopy with current and future ground based gravitational wave detectors over the next 10 years. For different population models, we estimate the likely number of events for which the subdominant mode can be detected and distinguished from the dominant mode. We show that black hole spectroscopy could significantly test general relativity for events seen by the proposed LIGO Voyager detectors.

gr-qc↗

arXiv2019

Detectability of the subdominant mode in a binary black hole ringdown

The ringdown is the late part of the post-merger signature emitted during the coalescence of two black holes and comprises of a superposition of quasi-normal-modes. Within general relativity, because of the no-hair theorems, the frequencies and damping times of these modes are entirely determined by the mass and angular momentum of the final Kerr black hole. A detection of multiple ringdown modes would potentially allow us to test the no-hair theorem from observational data. The parameters which determine whether sub-dominant ringdown modes can be detected are primarily the overall signal-to-noise ratio present in the ringdown signal, and on the amplitude of the subdominant mode with respect to the dominant mode. In this paper, we use Bayesian inference to determine the detectability of a subdominant mode in a set of simulated analytical ringdown signals. Focusing on the design sensitivity of the Advanced LIGO detectors, we systematically vary the signal-to-noise ratio of the ringdown signal, and the mode amplitude ratio in order to determine what kind of signals are promising for performing black hole spectroscopy.

gr-qc↗

arXiv2019

Potential Gravitational-wave and Gamma-ray Multi-messenger Candidate from Oct. 30, 2015

We present a search for binary neutron star mergers that produced gravitational-waves during the first observing run of Advanced LIGO and gamma-ray emission seen by either \textit{Swift}-BAT or Fermi-GBM, similar to GW170817 and GRB 170817A. We introduce a new method using a combined ranking statistic to detect sources that do not produce significant gravitational-wave or gamma-ray burst candidates individually. The current version of this search can increase by 70% the detections of joint gravitational-wave and gamma-ray signals. We find one possible candidate observed by LIGO and Fermi-GBM, 1-OGC 151030, at a false alarm rate of 1 in 13 years. If astrophysical, this candidate would correspond to a merger at $187^{+99}_{-87}\,$Mpc with source-frame chirp mass of $1.30^{+0.02}_{-0.03}\,\mathrm{M}_{\odot}$. If we assume the viewing angle must be $<30^{\circ}$ to be observed by \textit{Fermi}-GBM, our estimate of the distance would become $224^{+88}_{-78}\,$Mpc. By comparing the rate of binary neutron star mergers to our search-estimated rate of false alarms, we estimate that there is a 1 in 4 chance this candidate is astrophysical in origin.

astro-ph.HE↗

arXiv2019

Posterior samples of the parameters of binary black holes from Advanced LIGO, Virgo's second observing run

This paper presents a parameter estimation analysis of the seven binary black hole mergers---GW170104, GW170608, GW170729, GW170809, GW170814, GW170818, and GW170823---detected during the second observing run of the Advanced LIGO and Virgo observatories using the gravitational-wave open data. We describe the methodology for parameter estimation of compact binaries using gravitational-wave data, and we present the posterior distributions of the inferred astrophysical parameters. We release our samples of the posterior probability density function with tutorials on using and replicating our results presented in this paper.

gr-qc↗

arXiv2019

Investigating the noise residuals around the gravitational wave event GW150914

We use the Pearson cross-correlation statistic proposed by Liu and Jackson, and employed by Creswell et al., to look for statistically significant correlations between the LIGO Hanford and Livingston detectors at the time of the binary black hole merger GW150914. We compute this statistic for the calibrated strain data released by LIGO, using both the residuals provided by LIGO and using our own subtraction of a maximum-likelihood waveform that is constructed to model binary black hole mergers in general relativity. To assign a significance to the values obtained, we calculate the cross-correlation of both simulated Gaussian noise and data from the LIGO detectors at times during which no detection of gravitational waves has been claimed. We find that after subtracting the maximum likelihood waveform there are no statistically significant correlations between the residuals of the two detectors at the time of GW150914.

astro-ph.HE↗

arXiv2018

Parameter estimation for black hole echo signals and their statistical significance

Searching for black hole echo signals with gravitational waves provides a means of probing the near-horizon regime of these objects. We demonstrate a pipeline to efficiently search for these signals in gravitational wave data and calculate model selection probabilities between signal and no-signal hypotheses. As an example of its use we calculate Bayes factors for the Abedi-Dykaar-Afshordi (ADA) model on events in LIGO's first observing run and compare to existing results in the literature. We discuss the benefits of using a full likelihood exploration over existing search methods that used template banks and calculated p-values. We use the waveforms of ADA, although the method is easily extendable to other waveforms. With these waveforms we are able to demonstrate a range of echo amplitudes that is already is ruled out by the data.

gr-qc↗

arXiv2018

Observational tests of the black hole area increase law

The black hole area theorem implies that when two black holes merge, the area of the final black hole should be greater than the sum of the areas of the two original black holes. We examine how this prediction can be tested with gravitational-wave observations of binary black holes. By separately fitting the early inspiral and final ringdown stages, we calculate the posterior distributions for the masses and spins of the two initial and the final black holes. This yields posterior distributions for the change in the area and thus a statistical test of the validity of the area increase law. We illustrate this method with a GW150914-like binary black hole waveform calculated using numerical relativity, and detector sensitivities representative of both the first observing run and the design configuration of Advanced LIGO. We obtain a $\sim74.6\%$ probability that the simulated signal is consistent with the area theorem with current sensitivity, improving to $\sim99.9\%$ when Advanced LIGO reaches design sensitivity. An important ingredient in our test is a method of estimating when the post-merger signal is well-fit by a damped sinusoid ringdown waveform.

gr-qc↗

arXiv2018

PyCBC Inference: A Python-based parameter estimation toolkit for compact binary coalescence signals

We introduce new modules in the open-source PyCBC gravitational- wave astronomy toolkit that implement Bayesian inference for compact-object binary mergers. We review the Bayesian inference methods implemented and describe the structure of the modules. We demonstrate that the PyCBC Inference modules produce unbiased estimates of the parameters of a simulated population of binary black hole mergers. We show that the posterior parameter distributions obtained used our new code agree well with the published estimates for binary black holes in the first LIGO-Virgo observing run.

astro-ph.IM↗

arXiv2017

Minimum energy and the end of the inspiral in the post-Newtonian approximation

The early inspiral phase of a compact binary coalescence is well modelled by the post-Newtonian (PN) approximation to the orbital energy and gravitational wave flux. The transition from the inspiral phase to the plunge can be defined by the minimum energy circular orbit (MECO). In the extreme mass-ratio limit, up to the highest PN order known, the PN energy equals the energy of the exact Kerr solution. However, for comparable-mass systems the MECO of the PN energy does not exist when bodies have large spins. By including the exact Kerr limit and recently published post-Newtonian terms we extract a well-defined minimum of the orbital energy beyond which the plunge or merger occurs. We study the hybrid condition for a number of cases of both black hole and neutron stars and compare to other commonly employed definitions. Our method can be used for any known order of the post-Newtonian series and enables the MECO condition to be used to define the end of the inspiral phase for highly spinning, comparable mass systems.

gr-qc↗

arXiv2016

The PyCBC search for gravitational waves from compact binary coalescence

We describe the PyCBC search for gravitational waves from compact-object binary coalescences in advanced gravitational-wave detector data. The search was used in the first Advanced LIGO observing run and unambiguously identified two black hole binary mergers, GW150914 and GW151226. At its core, the PyCBC search performs a matched-filter search for binary merger signals using a bank of gravitational-wave template waveforms. We provide a complete description of the search pipeline including the steps used to mitigate the effects of noise transients in the data, identify candidate events and measure their statistical significance. The analysis is able to measure false-alarm rates as low as one per million years, required for confident detection of signals. Using data from initial LIGO's sixth science run, we show that the new analysis reduces the background noise in the search, giving a 30% increase in sensitive volume for binary neutron star systems over previous searches.

gr-qc↗