arXiv subjects

Joseph D. Romano

Publications and source records attributed to Joseph D. Romano.

At least 73 records · Page 4Linked to original sources

arXiv2021

PMLB v1.0: An open source dataset collection for benchmarking machine learning methods

Motivation: Novel machine learning and statistical modeling studies rely on standardized comparisons to existing methods using well-studied benchmark datasets. Few tools exist that provide rapid access to many of these datasets through a standardized, user-friendly interface that integrates well with popular data science workflows. Results: This release of PMLB provides the largest collection of diverse, public benchmark datasets for evaluating new machine learning and data science methods aggregated in one location. v1.0 introduces a number of critical improvements developed following discussions with the open-source community. Availability: PMLB is available at https://github.com/EpistasisLab/pmlb. Python and R interfaces for PMLB can be installed through the Python Package Index and Comprehensive R Archive Network, respectively.

cs.LG↗

arXiv2020

Common-spectrum process versus cross-correlation for gravitational-wave searches using pulsar timing arrays

The North American Nanohertz Observatory for Gravitational Waves (NANOGrav) has recently reported strong statistical evidence for a common-spectrum red-noise process for all pulsars, as seen in their 12.5-yr analysis for an isotropic stochastic gravitational-wave background. However, there is currently very little evidence for quadrupolar spatial correlations across the pulsars in the array, which is needed to make a confident claim of detection of a stochastic background. Here we give a frequentist analysis of a very simple signal+noise model showing that the current lack of evidence for spatial correlations is consistent with the magnitude of the correlation coefficients for pairs of Earth-pulsar baselines in the array, and the fact that pulsar timing arraysbare most-likely operating in the intermediate-signal regime. We derive analytic expressions that allow one to compare the expected values of the signal-to-noise ratios for both the common-spectrum and cross-correlation estimators.

gr-qc↗

arXiv2020

Frequentist versus Bayesian analyses: Cross-correlation as an (approximate) sufficient statistic for LIGO-Virgo stochastic background searches

Sufficient statistics are combinations of data in terms of which the likelihood function can be rewritten without loss of information. Depending on the data volume reduction, the use of sufficient statistics as a preliminary step in a Bayesian analysis can lead to significant increases in efficiency when sampling from posterior distributions of model parameters. Here we show that the frequency integrand of the cross-correlation statistic and its variance are approximate sufficient statistics for ground-based searches for stochastic gravitational-wave backgrounds. The sufficient statistics are approximate because one works in the weak-signal approximation and uses measured estimates of the auto-correlated power in each detector. Using analytic and numerical calculations, we prove that LIGO-Virgo's hybrid frequentist-Bayesian parameter estimation analysis is equivalent to a fully Bayesian analysis. This work closes a gap in the LIGO-Virgo literature, and suggests directions for additional searches.

gr-qc↗

arXiv2020

Model Dependence of Bayesian Gravitational-Wave Background Statistics for Pulsar Timing Arrays

Pulsar timing array (PTA) searches for a gravitational-wave background (GWB) typically include time-correlated "red" noise models intrinsic to each pulsar. Using a simple simulated PTA dataset with an injected GWB signal we show that the details of the red noise models used, including the choice of amplitude priors and even which pulsars have red noise, have a striking impact on the GWB statistics, including both upper limits and estimates of the GWB amplitude. We find that the standard use of uniform priors on the red noise amplitude leads to 95% upper limits, as calculated from one-sided Bayesian credible intervals, that are less than the injected GWB amplitude 50% of the time. In addition, amplitude estimates of the GWB are systematically lower than the injected value by 10-40%, depending on which models and priors are chosen for the intrinsic red noise. We tally the effects of model and prior choice and demonstrate how a "dropout" model, which allows flexible use of red noise models in a Bayesian approach, can improve GWB estimates throughout.

astro-ph.IM↗

arXiv2020

Is deep learning necessary for simple classification tasks?

Automated machine learning (AutoML) and deep learning (DL) are two cutting-edge paradigms used to solve a myriad of inductive learning tasks. In spite of their successes, little guidance exists for when to choose one approach over the other in the context of specific real-world problems. Furthermore, relatively few tools exist that allow the integration of both AutoML and DL in the same analysis to yield results combining both of their strengths. Here, we seek to address both of these issues, by (1.) providing a head-to-head comparison of AutoML and DL in the context of binary classification on 6 well-characterized public datasets, and (2.) evaluating a new tool for genetic programming-based AutoML that incorporates deep estimators. Our observations suggest that AutoML outperforms simple DL classifiers when trained on similar datasets for binary classification but integrating DL into AutoML improves classification performance even further. However, the substantial time needed to train AutoML+DL pipelines will likely outweigh performance advantages in many applications.

cs.LG↗

arXiv2020

Estimating the angular power spectrum of the gravitational-wave background in the presence of shot noise

There has been much recent interest in studying anisotropies in the astrophysical gravitational-wave (GW) background, as these could provide us with interesting new information about galaxy clustering and large-scale structure. However, this information is obscured by shot noise, caused by the finite number of GW sources that contribute to the background at any given time. We develop a new method for estimating the angular spectrum of anisotropies, based on the principle of combining statistically-independent data segments. We show that this gives an unbiased estimate of the true, astrophysical spectrum, removing the offset due to shot noise power, and that in the limit of many data segments, it is the most efficient (i.e. lowest-variance) estimator possible.

astro-ph.CO↗

arXiv2019

Realistic sensitivity curves for pulsar timing arrays

We construct realistic sensitivity curves for pulsar timing array searches for gravitational waves, incorporating both red and white noise contributions to individual pulsar noise spectra, as well as the effect of fitting to a pulsar timing model. We demonstrate the method on both simulated pulsars and a realistic array consisting of a subset of NANOGrav pulsars used in recent analyses. A comparison between the results presented here and exact sensitivity curves shows agreement to tens of percent. The resulting sensitivity curves can be used to assess the detectability of predicted gravitational-wave signals in the nanohertz frequency band in a fraction of the time that it would take to compute the exact sensitivity curves.

gr-qc↗

arXiv2018

An Acoustical Analogue of a Galactic-scale Gravitational-Wave Detector

By precisely monitoring the "ticks" of Nature's most precise clocks (millisecond pulsars), scientists are trying to detect the "ripples in spacetime" (gravitational waves) produced by the inspirals of supermassive black holes in the centers of distant merging galaxies. Here we describe a relatively simple demonstration that uses two metronomes and a microphone to illustrate several techniques used by pulsar astronomers to search for gravitational waves. An adapted version of this demonstration could be used as an instructional laboratory investigation at the undergraduate level.

physics.ed-ph↗

arXiv2017

Detection methods for stochastic gravitational-wave backgrounds: a unified treatment

We review detection methods that are currently in use or have been proposed to search for a stochastic background of gravitational radiation. We consider both Bayesian and frequentist searches using ground-based and space-based laser interferometers, spacecraft Doppler tracking, and pulsar timing arrays; and we allow for anisotropy, non-Gaussianity, and non-standard polarization states. Our focus is on relevant data analysis issues, and not on the particular astrophysical or early Universe sources that might give rise to such backgrounds. We provide a unified treatment of these searches at the level of detector response functions, detection sensitivity curves, and, more generally, at the level of the likelihood function, since the choice of signal and noise models and prior probability distributions are actually what define the search. Pedagogical examples are given whenever possible to compare and contrast different approaches. We have tried to make the article as self-contained and comprehensive as possible, targeting graduate students and new researchers looking to enter this field.

gr-qc↗

arXiv2015

Mapping gravitational-wave backgrounds in modified theories of gravity using pulsar timing arrays

We extend our previous work on applying CMB techniques to the mapping of gravitational-wave backgrounds to backgrounds which have non-GR polarisations. Our analysis and results are presented in the context of pulsar-timing array observations, but the overarching methods are general, and can be easily applied to LIGO or eLISA observations using appropriately modified response functions. Analytic expressions for the pulsar-timing response to gravitational waves with non-GR polarisation are given for each mode of a spin-weighted spherical-harmonic decomposition of the background, which permit the signal to be mapped across the sky to any desired resolution. We also derive the pulsar-timing overlap reduction functions for the various non-GR polarisations, finding analytic forms for anisotropic backgrounds with scalar-transverse ("breathing") and vector-longitudinal polarisations, and a semi-analytic form for scalar-longitudinal backgrounds. Our results indicate that pulsar-timing observations will be completely insensitive to scalar-transverse mode anisotropies in the polarisation amplitude beyond dipole, and anisotropies in the power beyond quadrupole. Analogously to our previous findings that pulsar-timing observations lack sensitivity to tensor-curl modes for a transverse-traceless tensor background, we also find insensitivity to vector-curl modes for a vector-longitudinal background.

gr-qc↗

arXiv2015

A Mock Data and Science Challenge for Detecting an Astrophysical Stochastic Gravitational-Wave Background with Advanced LIGO and Advanced Virgo

The purpose of this mock data and science challenge is to prepare the data analysis and science interpretation for the second generation of gravitational-wave experiments Advanced LIGO-Virgo in the search for a stochastic gravitational-wave background signal of astrophysical origin. Here we present a series of signal and data challenges, with increasing complexity, whose aim is to test the ability of current data analysis pipelines at detecting an astrophysically produced gravitational-wave background, test parameter estimation methods and interpret the results. We introduce the production of these mock data sets that includes a realistic observing scenario data set where we account for different sensitivities of the advanced detectors as they are continuously upgraded toward their design sensitivity. After analysing these with the standard isotropic cross-correlation pipeline we find that we are able to recover the injected gravitational-wave background energy density to within $2σ$ for all of the data sets and present the results from the parameter estimation. The results from this mock data and science challenge show that advanced LIGO and Virgo will be ready and able to make a detection of an astrophysical gravitational-wave background within a few years of operations of the advanced detectors, given a high enough rate of compact binary coalescing events.

astro-ph.HE↗

arXiv2015

When is a gravitational-wave signal stochastic?

We discuss the detection of gravitational-wave backgrounds in the context of Bayesian inference and suggest a practical definition of what it means for a signal to be considered stochastic---namely, that the Bayesian evidence favors a stochastic signal model over a deterministic signal model. A signal can further be classified as Gaussian-stochastic if a Gaussian signal model is favored. In our analysis we use Bayesian model selection to choose between several signal and noise models for simulated data consisting of uncorrelated Gaussian detector noise plus a superposition of sinusoidal signals from an astrophysical population of gravitational-wave sources. For simplicity, we consider co-located and co-aligned detectors with white detector noise, but the method can be extended to more realistic detector configurations and power spectra. The general trend we observe is that a deterministic model is favored for small source numbers, a non-Gaussian stochastic model is preferred for intermediate source numbers, and a Gaussian stochastic model is preferred for large source numbers. However, there is very large variation between individual signal realizations, leading to fuzzy boundaries between the three regimes. We find that a hybrid, trans-dimensional model comprised of a deterministic signal model for individual bright sources and a Gaussian-stochastic signal model for the remaining confusion background outperforms all other models in most instances.

gr-qc↗

arXiv2015

Phase-coherent mapping of gravitational-wave backgrounds using ground-based laser interferometers

We extend the formalisms developed in Gair et al. and Cornish and van Haasteren to create maps of gravitational-wave backgrounds using a network of ground-based laser interferometers. We show that in contrast to pulsar timing arrays, which are insensitive to half of the gravitational-wave sky (the curl modes), a network of ground-based interferometers is sensitive to both the gradient and curl components of the background. The spatial separation of a network of interferometers, or of a single interferometer at different times during its rotational and orbital motion around the Sun, allows for recovery of both components. We derive expressions for the response functions of a laser interferometer in the small-antenna limit, and use these expressions to calculate the overlap reduction function for a pair of interferometers. We also construct maximum-likelihood estimates of the + and x-polarization modes of the gravitational-wave sky in terms of the response matrix for a network of ground-based interferometers, evaluated at discrete times during Earth's rotational and orbital motion around the Sun. We demonstrate the feasibility of this approach for some simple simulated backgrounds (a single point source and spatially-extended distributions having only grad or curl components), calculating maximum-likelihood sky maps and uncertainty maps based on the (pseudo)inverse of the response matrix. The distinction between this approach and standard methods for mapping gravitational-wave power is also discussed.

gr-qc↗

arXiv2015

Understanding the gravitational-wave Hellings and Downs curve for pulsar timing arrays in terms of sound and electromagnetic waves

Searches for stochastic gravitational-wave backgrounds using pulsar timing arrays look for correlations in the timing residuals induced by the background across the pulsars in the array. The correlation signature of an isotropic, unpolarized gravitational-wave background predicted by general relativity follows the so-called Hellings and Downs curve, which is a relatively simple function of the angle between a pair of Earth-pulsar baselines. In this paper, we give a pedagogical discussion of the Hellings and Downs curve for pulsar timing arrays, considering simpler analogous scenarios involving sound and electromagnetic waves. We calculate Hellings-and-Downs-type functions for these two scenarios and develop a framework suitable for doing more general correlation calculations.

gr-qc↗

arXiv2014

Time-domain Implementation of the Optimal Cross-Correlation Statistic for Stochastic Gravitational-Wave Background Searches in Pulsar Timing Data

Supermassive black hole binaries, cosmic strings, relic gravitational waves from inflation, and first order phase transitions in the early universe are expected to contribute to a stochastic background of gravitational waves in the 10^(-9) Hz-10^(-7) Hz frequency band. Pulsar timing arrays (PTAs) exploit the high precision timing of radio pulsars to detect signals at such frequencies. Here we present a time-domain implementation of the optimal cross-correlation statistic for stochastic background searches in PTA data. Due to the irregular sampling typical of PTA data as well as the use of a timing model to predict the times-of-arrival of radio pulses, time-domain methods are better suited for gravitational wave data analysis of such data. We present a derivation of the optimal cross-correlation statistic starting from the likelihood function, a method to produce simulated stochastic background signals, and a rigorous derivation of the scaling laws for the signal-to-noise ratio of the cross-correlation statistic in the two relevant PTA regimes: the weak signal limit where instrumental noise dominates over the gravitational wave signal at all frequencies, and a second regime where the gravitational wave signal dominates at the lowest frequencies.

astro-ph.IM↗

arXiv2014

Mapping gravitational-wave backgrounds using methods from CMB analysis: Application to pulsar timing arrays

We describe an alternative approach to the analysis of gravitational-wave backgrounds, based on the formalism used to characterise the polarisation of the cosmic microwave background. In contrast to standard analyses, this approach makes no assumptions about the nature of the background and so has the potential to reveal much more about the physical processes that generated it. An arbitrary background can be decomposed into modes whose angular dependence on the sky is given by gradients and curls of spherical harmonics. We derive the pulsar timing overlap reduction functions for the individual modes, which are given by simple combinations of spherical harmonics evaluated at the pulsar locations. We show how these can be used to recover the components of an arbitrary background, giving explicit results for both isotropic and anisotropic uncorrelated backgrounds. We also find that the response of a pulsar timing array to curl modes is identically zero, so half of the gravitational-wave sky will never be observed using pulsar timing, no matter how many pulsars are included in the array. An isotropic, unpolarised and uncorrelated background can be accurately represented using only three modes, and so a search of this type will be only slightly more complicated than the standard cross-correlation search using the Hellings and Downs overlap reduction function. However, by measuring the components of individual modes of the background and checking for consistency with isotropy, this approach has the potential to reveal much more information. Each individual mode on its own describes a background that is correlated between different points on the sky. A measurement of the components that indicates the presence of correlations in the background on large angular scales would suggest startling new physics.

gr-qc↗

arXiv2013

Towards a unified treatment of gravitational-wave data analysis

We present a unified description of gravitational-wave data analysis that unites the template-based analysis used to detect deterministic signals from well-modeled sources, such as binary-black-hole mergers, with the cross-correlation analysis used to detect stochastic gravitational-wave backgrounds. We also discuss the connection between template-based analyses and those that target poorly-modeled bursts of gravitational waves, and suggest a new approach for detecting burst signals.

gr-qc↗

arXiv2013

The stochastic background: scaling laws and time to detection for pulsar timing arrays

We derive scaling laws for the signal-to-noise ratio of the optimal cross-correlation statistic, and show that the large power-law increase of the signal-to-noise ratio as a function of the the observation time $T$ that is usually assumed holds only at early times. After enough time has elapsed, pulsar timing arrays enter a new regime where the signal to noise only scales as $\sqrt{T}$. In addition, in this regime the quality of the pulsar timing data and the cadence become relatively un-important. This occurs because the lowest frequencies of the pulsar timing residuals become gravitational-wave dominated. Pulsar timing arrays enter this regime more quickly than one might naively suspect. For T=10 yr observations and typical stochastic background amplitudes, pulsars with residual RMSs of less than about $1\,μ$s are already in that regime. The best strategy to increase the detectability of the background in this regime is to increase the number of pulsars in the array. We also perform realistic simulations of the NANOGrav pulsar timing array, which through an aggressive pulsar survey campaign adds new millisecond pulsars regularly to its array, and show that a detection is possible within a decade, and could occur as early as 2016.

astro-ph.IM↗