arXiv subjects

Jonathan Gair

Publications and source records attributed to Jonathan Gair.

At least 55 records · Page 3Linked to original sources

arXiv2023

Bayesian analysis of systematic errors in the determination of the constant of gravitation

Measurements of the gravitational constant $G$ are notoriously difficult. Individual state-of-the-art experiments have managed to determine the value of $G$ with high precision: although, when considered collectively, the range in the measured values of $G$ far exceeds individual uncertainties, suggesting the presence of unaccounted for systematic effects. Here, we propose a Bayesian framework to account for the presence of systematic errors in the various measurement of $G$ while proposing a consensus value, following two paths: a parametric approach, based on the Maximum Entropy Principle, and a non-parametric one, the latter being a very flexible approach not committed to any specific functional form. With both our methods, we find that the uncertainty on this fundamental constant, once systematics are included, is significantly larger than what quoted in CODATA 2018. Moreover, the morphology of the non-parametric distribution hints towards the presence of several sources of unaccounted for systematics. In light of this, we recommend a consensus value for the gravitational constant $G = 6.6740^{+0.0015}_{-0.0015} \times 10^{-11}\ \mathrm{m}^3\ \mathrm{kg}^{-1}\ \mathrm{s}^{-2}$.

gr-qc↗

arXiv2023

Impact of the noise knowledge uncertainty for the science exploitation of cosmological and astrophysical stochastic gravitational wave background with LISA

This paper investigates the impact of a lack of knowledge of the instrumental noise on the characterisation of stochastic gravitational wave backgrounds with the Laser Interferometer Space Antenna (LISA). We focus on constraints on modelled backgrounds that represent the possible backgrounds from the mergers of binary black holes of stellar origin, from primordial black hole generation, from non-standard inflation, and from sound wave production during cosmic fluid phase transitions. We use splines to model generic, slowly varying, uncertainties in the auto and cross-spectral densities of the LISA time delay interferometry channels. We find that allowing for noise knowledge uncertainty in this way leads to one to two orders of magnitude degradation in our ability to constrain stochastic backgrounds, and a corresponding increase in the background energy density required for a confident detection. We also find that to avoid this degradation, the LISA noise would have to be known at the sub-percent level, which is unlikely to be achievable in practice.

gr-qc↗

arXiv2023

Group equivariant neural posterior estimation

Simulation-based inference with conditional neural density estimators is a powerful approach to solving inverse problems in science. However, these methods typically treat the underlying forward model as a black box, with no way to exploit geometric properties such as equivariances. Equivariances are common in scientific models, however integrating them directly into expressive inference networks (such as normalizing flows) is not straightforward. We here describe an alternative method to incorporate equivariances under joint transformations of parameters and data. Our method -- called group equivariant neural posterior estimation (GNPE) -- is based on self-consistently standardizing the "pose" of the data while estimating the posterior over parameters. It is architecture-independent, and applies both to exact and approximate equivariances. As a real-world application, we use GNPE for amortized inference of astrophysical binary black hole systems from gravitational-wave observations. We show that GNPE achieves state-of-the-art accuracy while reducing inference times by three orders of magnitude.

cs.LG↗

arXiv2023

Real-time gravitational-wave science with neural posterior estimation

We demonstrate unprecedented accuracy for rapid gravitational-wave parameter estimation with deep learning. Using neural networks as surrogates for Bayesian posterior distributions, we analyze eight gravitational-wave events from the first LIGO-Virgo Gravitational-Wave Transient Catalog and find very close quantitative agreement with standard inference codes, but with inference times reduced from O(day) to a minute per event. Our networks are trained using simulated data, including an estimate of the detector-noise characteristics near the event. This encodes the signal and noise models within millions of neural-network parameters, and enables inference for any observed data consistent with the training distribution, accounting for noise nonstationarity from event to event. Our algorithm -- called "DINGO" -- sets a new standard in fast-and-accurate inference of physical parameters of detected gravitational-wave events, which should enable real-time data analysis without sacrificing accuracy.

gr-qc↗

arXiv2023

Neural Importance Sampling for Rapid and Reliable Gravitational-Wave Inference

We combine amortized neural posterior estimation with importance sampling for fast and accurate gravitational-wave inference. We first generate a rapid proposal for the Bayesian posterior using neural networks, and then attach importance weights based on the underlying likelihood and prior. This provides (1) a corrected posterior free from network inaccuracies, (2) a performance diagnostic (the sample efficiency) for assessing the proposal and identifying failure cases, and (3) an unbiased estimate of the Bayesian evidence. By establishing this independent verification and correction mechanism we address some of the most frequent criticisms against deep learning for scientific inference. We carry out a large study analyzing 42 binary black hole mergers observed by LIGO and Virgo with the SEOBNRv4PHM and IMRPhenomXPHM waveform models. This shows a median sample efficiency of $\approx 10\%$ (two orders-of-magnitude better than standard samplers) as well as a ten-fold reduction in the statistical uncertainty in the log evidence. Given these advantages, we expect a significant impact on gravitational-wave inference, and for this approach to serve as a paradigm for harnessing deep learning methods in scientific applications.

gr-qc↗

arXiv2022

Aberration of gravitational waveforms by peculiar velocity

One key prediction of General Relativity is that gravitational waves are emitted with a pure spin-2 polarisation. Any extra polarisation mode, spin-1 or spin-0, is consequently considered a smoking gun for deviations from General Relativity. In this paper, we show that the velocity of merging binaries with respect to the observer gives rise to spin-1 polarisation in the observer frame even in the context of General Relativity. These are pure projection effects, proportional to the plus and cross polarisations in the source frame, hence they do not correspond to new degrees of freedom. We demonstrate that the spin-1 modes can always be rewritten as pure spin-2 modes coming from an aberrated direction. Since gravitational waves are not isotropically emitted around binary systems, this aberration modifies the apparent orientation of the binary system with respect to the observer: the system appears slightly rotated due to the source velocity. Fortunately, this bias does not propagate to other parameters of the system (and therefore does not spoil tests of General Relativity), since the impact of the velocity can be fully reabsorbed into new orientation angles.

gr-qc↗

arXiv2022

Adapting to noise distribution shifts in flow-based gravitational-wave inference

Deep learning techniques for gravitational-wave parameter estimation have emerged as a fast alternative to standard samplers $\unicode{x2013}$ producing results of comparable accuracy. These approaches (e.g., DINGO) enable amortized inference by training a normalizing flow to represent the Bayesian posterior conditional on observed data. By conditioning also on the noise power spectral density (PSD) they can even account for changing detector characteristics. However, training such networks requires knowing in advance the distribution of PSDs expected to be observed, and therefore can only take place once all data to be analyzed have been gathered. Here, we develop a probabilistic model to forecast future PSDs, greatly increasing the temporal scope of DINGO networks. Using PSDs from the second LIGO-Virgo observing run (O2) $\unicode{x2013}$ plus just a single PSD from the beginning of the third (O3) $\unicode{x2013}$ we show that we can train a DINGO network to perform accurate inference throughout O3 (on 37 real events). We therefore expect this approach to be a key component to enable the use of deep learning techniques for low-latency analyses of gravitational waves.

gr-qc↗

arXiv2022

New Horizons for Fundamental Physics with LISA

The Laser Interferometer Space Antenna (LISA) has the potential to reveal wonders about the fundamental theory of nature at play in the extreme gravity regime, where the gravitational interaction is both strong and dynamical. In this white paper, the Fundamental Physics Working Group of the LISA Consortium summarizes the current topics in fundamental physics where LISA observations of GWs can be expected to provide key input. We provide the briefest of reviews to then delineate avenues for future research directions and to discuss connections between this working group, other working groups and the consortium work package teams. These connections must be developed for LISA to live up to its science potential in these areas.

gr-qc↗

arXiv2022

Known unknowns: assessing the impact of instrumental calibration uncertainty on LISA science

The primary scientific results of the future space-based gravitational wave interferometer LISA will come from the parameter inference of a large variety of gravitational wave sources. However, the presence of calibration errors could potentially degrade the measurement precision of the system parameters. Here, we assess the impact of calibration uncertainties on parameter estimation for individual sources, focusing on massive black holes, extreme-mass-ratio inspirals (EMRIs), galactic binaries, and stellar origin black hole binaries. Using a Fisher matrix formalism, we investigate how the measurement precision of source parameters degrades as a function of the size of the assumed calibration uncertainties. If we require that parameter measurements are degraded by no more than a factor of two relative to their value in the absence of calibration error, we find that calibration errors should be smaller than a few tenths of a percent in amplitude and $10^{-3}$ in phase. We also investigate the possibility of using verification binaries and EMRIs to constrain calibration uncertainties. Verification binaries can constrain amplitude calibration uncertainties at the level of a few percent, while both source types can provide constrain phase calibration at the level of a few$\times10^{-2}$.

gr-qc↗

arXiv2022

Constraining Unmodeled Physics with Compact Binary Mergers from GWTC-1

We present a flexible model to describe the effects of generic deviations of observed gravitational wave signals from modeled waveforms in the LIGO and Virgo gravitational wave detectors. With the detection of 11 gravitational wave events from the GWTC-1 catalog, we are able to constrain possible deviations from our modeled waveforms. In this paper we present our coherent spline model that describes the deviations, then choose to validate our model on an example phenomenological and astrophysically motivated departure in waveforms based on extreme spontaneous scalarization. We find that the model is capable of recovering the simulated deviations. By performing model comparisons we observe that the spline model effectively describes the simulated departures better than a normal compact binary coalescence (CBC) model. We analyze the entire GWTC-1 catalog of events with our model and compare it to a normal CBC model, finding that there are no significant departures from the modeled template gravitational waveforms used.

gr-qc↗

arXiv2021

Gravitational wave cosmology with extreme mass-ratio inspirals

The Laser Interferometer Space Antenna (LISA) will open the mHz frequency window of the gravitational wave (GW) landscape. Among all the new GW sources expected to emit in this frequency band, extreme mass-ratio inspirals (EMRIs) constitute a unique laboratory for astrophysics and fundamental physics. Here we show that EMRIs can also be used to extract relevant cosmological information, complementary to both electromagnetic (EM) and other GW observations. By using the loudest EMRIs (SNR$>$100) detected by LISA as dark standard sirens, statistically matching their sky localisation region with mock galaxy catalogs, we find that constraints on $H_0$ can reach $\sim$1.1% ($\sim$3.6%) accuracy, at the 90% credible level, in our best (worst) case scenario. By considering a dynamical dark energy (DE) cosmological model, with $Λ$CDM parameters fixed by other observations, we further show that in our best (worst) case scenario $\sim$5.9% ($\sim$12.3%) relative uncertainties at the 90% credible level can be obtained on $w_0$, the DE equation of state parameter. Besides being relevant in their own right, EMRI measurements will be affected by different systematics compared to both EM and ground-based GW observations. Cross validation with complementary cosmological measurements will therefore be of paramount importance, especially if convincing evidence of physics beyond $Λ$CDM emerges from future observations.

astro-ph.CO↗

arXiv2021

Observing intermediate-mass black holes and the upper--stellar-mass gap with LIGO and Virgo

Using ground-based gravitational-wave detectors, we probe the mass function of intermediate-mass black holes (IMBHs) wherein we also include BHs in the upper mass gap $\sim 60-130~M_\odot$. Employing the projected sensitivity of the upcoming LIGO and Virgo fourth observing (O4) run, we perform Bayesian analysis on quasi-circular non-precessing, spinning IMBH binaries (IMBHBs) with total masses $50\mbox{--} 500\, M_\odot$, mass ratios 1.25, 4, and 10, and dimensionless spins up to 0.95, and estimate the precision with which the source-frame parameters can be measured. We find that, at $2σ$, the mass of the heavier component of IMBHBs can be constrained with an uncertainty of $\sim 10-40\%$ at a signal-to-noise ratio of $20$. Focusing on the stellar-mass gap with new tabulations of the $^{12}\text{C}(α, γ)^{16} \text{O}$ reaction rate and its uncertanties, we evolve massive helium core stars using \MESA\, to establish the lower and upper edge of the mass gap as $\simeq$\,59$^{+34}_{-13}$\,$M_{\odot}$ and $\simeq$\,139$^{+30}_{-14}$\,$M_{\odot}$ respectively, where the error bars give the mass range that follows from the $\pm 3σ$ uncertainty in the $^{12}\text{C}(α, γ) ^{16} \text{O}$ nuclear reaction rate. We find that high resolution of the tabulated reaction rate and fine temporal resolution are necessary to resolve the peak of the BH mass spectrum. We then study IMBHBs with components lying in the mass gap and show that the O4 run will be able to robustly identify most such systems. Finally, we re-analyse GW190521 with a state-of-the-art aligned-spin waveform model, finding that the primary mass lies in the mass gap with 90\% credibility.

gr-qc↗

arXiv2021

Mapping the inhomogeneous Universe with Standard Sirens: Degeneracy between inhomogeneity and modified gravity theories

The detection of gravitational waves (GWs) and an accompanying electromagnetic (E/M) counterpart have been suggested as a future probe for cosmology and theories of gravity. In this paper, we present calculations of the luminosity distance of sources taking into account inhomogeneities in the matter distribution that are predicted in numerical simulations of structure formation. In addition, we show that inhomogeneities resulting from clustering of matter can mimic certain classes of modified gravity theories, or other effects that dampen GW amplitudes, and deviations larger than $δν\sim \mathcal{O}(0.1)\ (99\%\ \rm{C.L.})$ to the extra friction term $ν$, from zero, would be necessary to distinguish them. For these, we assume mock GWs sources, with known redshift, based on binary population synthesis models, between redshifts $z=0$ and $z=5$. We show that future GW detectors, like Einstein Telescope or Cosmic Explorer, will be needed for strong constraints on the inhomogeneity parameters and breaking the degeneracy between modified gravity effects and matter anisotropies by measuring $ν$ at $5 \%$ and $1 \%$ level with $100$ and $350$ events respectively.

astro-ph.CO↗

arXiv2020

Identifying and Addressing Nonstationary LISA Noise

We anticipate noise from the Laser Interferometer Space Antenna (LISA) will exhibit nonstationarities throughout the duration of its mission due to factors such as antenna repointing, cyclostationarities from spacecraft motion, and glitches as highlighted by LISA Pathfinder. In this paper, we use a surrogate data approach to test the stationarity of a time series which does not rely on the Gaussianity assumption. The main goal is to identify noise nonstationarities in the future LISA mission. This will be necessary for determining how often the LISA noise power spectral density (PSD) will need to be updated for parameter estimation routines. We conduct a thorough simulation study illustrating the power/size of various versions of the hypothesis tests, and then apply these approaches to differential acceleration measurements from LISA Pathfinder. We also develop a data analysis strategy for addressing nonstationarities in the LISA PSD, where we update the noise PSD over time, while simultaneously conducting parameter estimation, with a focus on planned data gaps.

gr-qc↗

arXiv2020

Complete parameter inference for GW150914 using deep learning

The LIGO and Virgo gravitational-wave observatories have detected many exciting events over the past five years. As the rate of detections grows with detector sensitivity, this poses a growing computational challenge for data analysis. With this in mind, in this work we apply deep learning techniques to perform fast likelihood-free Bayesian inference for gravitational waves. We train a neural-network conditional density estimator to model posterior probability distributions over the full 15-dimensional space of binary black hole system parameters, given detector strain data from multiple detectors. We use the method of normalizing flows---specifically, a neural spline normalizing flow---which allows for rapid sampling and density estimation. Training the network is likelihood-free, requiring samples from the data generative process, but no likelihood evaluations. Through training, the network learns a global set of posteriors: it can generate thousands of independent posterior samples per second for any strain data consistent with the prior and detector noise characteristics used for training. By training with the detector noise power spectral density estimated at the time of GW150914, and conditioning on the event strain data, we use the neural network to generate accurate posterior samples consistent with analyses using conventional sampling techniques.

astro-ph.IM↗

arXiv2020

Gravitational-wave parameter estimation with autoregressive neural network flows

We introduce the use of autoregressive normalizing flows for rapid likelihood-free inference of binary black hole system parameters from gravitational-wave data with deep neural networks. A normalizing flow is an invertible mapping on a sample space that can be used to induce a transformation from a simple probability distribution to a more complex one: if the simple distribution can be rapidly sampled and its density evaluated, then so can the complex distribution. Our first application to gravitational waves uses an autoregressive flow, conditioned on detector strain data, to map a multivariate standard normal distribution into the posterior distribution over system parameters. We train the model on artificial strain data consisting of IMRPhenomPv2 waveforms drawn from a five-parameter $(m_1, m_2, ϕ_0, t_c, d_L)$ prior and stationary Gaussian noise realizations with a fixed power spectral density. This gives performance comparable to current best deep-learning approaches to gravitational-wave parameter estimation. We then build a more powerful latent variable model by incorporating autoregressive flows within the variational autoencoder framework. This model has performance comparable to Markov chain Monte Carlo and, in particular, successfully models the multimodal $ϕ_0$ posterior. Finally, we train the autoregressive latent variable model on an expanded parameter space, including also aligned spins $(χ_{1z}, χ_{2z})$ and binary inclination $θ_{JN}$, and show that all parameters and degeneracies are well-recovered. In all cases, sampling is extremely fast, requiring less than two seconds to draw $10^4$ posterior samples.

astro-ph.IM↗

arXiv2020

A Precessing Numerical Relativity Waveform Surrogate Model for Binary Black Holes: A Gaussian Process Regression Approach

Gravitational wave astrophysics relies heavily on the use of matched filtering both to detect signals in noisy data from detectors, and to perform parameter estimation on those signals. Matched filtering relies upon prior knowledge of the signals expected to be produced by a range of astrophysical systems, such as binary black holes. These waveform signals can be computed using numerical relativity techniques, where the Einstein field equations are solved numerically, and the signal is extracted from the simulation. Numerical relativity simulations are, however, computationally expensive, leading to the need for a surrogate model which can predict waveform signals in regions of the physical parameter space which have not been probed directly by simulation. We present a method for producing such a surrogate using Gaussian process regression which is trained directly on waveforms generated by numerical relativity. This model returns not just a single interpolated value for the waveform at a new point, but a full posterior probability distribution on the predicted value. This model is therefore an ideal component in a Bayesian analysis framework, through which the uncertainty in the interpolation can be taken into account when performing parameter estimation of signals.

gr-qc↗

arXiv2019

Astrometric Effects of Gravitational Wave Backgrounds with non-Luminal Propagation Speeds

A passing gravitational wave causes a deflection in the apparent astrometric positions of distant stars. The effect of the speed of the gravitational wave on this astrometric shift is discussed. A stochastic background of gravitational waves would result in a pattern of astrometric deflections which are correlated on large angular scales. These correlations are quantified and investigated for backgrounds of gravitational waves with sub- and super-luminal group velocities. The statistical properties of the correlations are depicted in two equivalent and related ways: as correlation curves and as angular power spectra. Sub-(super-)luminal gravitational wave backgrounds have the effect of enhancing (suppressing) the power in low-order angular modes. Analytical representations of the redshift-redshift and redshift-astrometry correlations are also derived. The potential for using this effect for constraining the speed of gravity is discussed.

gr-qc↗