arXiv subjects

Sanjit Mitra

Publications and source records attributed to Sanjit Mitra.

At least 37 records · Page 2Linked to original sources

arXiv2021

Improving significance of binary black hole mergers in Advanced LIGO data using deep learning : Confirmation of GW151216

We present a novel Machine Learning (ML) based strategy to search for binary black hole (BBH) mergers in data from ground-based gravitational wave (GW) observatories. This is the first ML-based search that not only recovers all the compact binary coalescences (CBCs) in the first GW transients catalog (GWTC-1), but also makes a clean detection of GW151216 by only adding a new coincident ranking statistic (MLStat) to a standard analysis that was used for GWTC-1. In CBC searches, reducing contamination by terrestrial and instrumental transients, which create a loud noise background by triggering numerous false alarms, is crucial to improving the sensitivity for detecting true events. The sheer volume of data and a large number of expected detections also prompts the use of ML techniques. We perform transfer learning to train "InceptionV3", a pre-trained deep neural network, along with curriculum learning to distinguish GW signals from noisy events by analysing their continuous wavelet transform (CWT) maps. MLStat incorporates information from this ML classifier into the coincident search likelihood used by the standard PyCBC search. This leads to at least an order of magnitude improvement in the inverse false-alarm-rate (IFAR) for the previously "low significance" events GW151012, GW170729 and GW151216. We also perform the parameter estimation of GW151216 using SEOBNRv4HM_ROM. We carry out an injection study to show that MLStat brings substantial improvement to the detection sensitivity of Advanced LIGO for all compact binary coalescences. The average improvement in the sensitive volume is ~10% for low chirp masses (0.8-5 Msun), and ~30% for higher masses (5-50 Msun). This work demonstrates the immense potential and readiness of MLStat for finding new sources in current data and the possibility of its adaptation in similar searches.

gr-qc↗

arXiv2021

Unified mapmaking for an anisotropic stochastic gravitational wave background

A stochastic gravitational wave background (SGWB), created by the superposition of signals from unresolved astrophysical sources, may be detected in the next few years. Estimating the variation of intensity across the sky can, therefore, play a key role in improving our understanding of astrophysical models. Skymaps have been produced for all the data-taking runs of the advanced ground-based interferometric detectors. While these maps are being produced in pixel and SpH bases, to probe, respectively, localized and diffuse astrophysical and cosmological sources, with algorithms that employ cross-correlation as the common strategy, the underlying algebra and numerical implementation remain different. As a consequence, there was a need for producing skymaps in both bases in those analyses. We show that these manifestly redundant methods could indeed be unified to a single analysis that can probe very different scales and demonstrate it by applying them on real data. We first develop the algebra to show that the results in two different bases are easily transformable. We then incorporate both the schemes in the now-standard analysis pipeline for anisotropic SGWB, PyStoch. This will enable SGWB anisotropy searches in SpH basis also to take full advantage of integrated \hpx tools and makes it computationally feasible to perform the search in every frequency bin. We, however, follow a different approach for direct estimation of the SpH moments. We show that the results obtained from these different methods match very well; the differences are less than $0.1$\% for the SpH moments and less than $0.01$\% for the Fisher information matrices. Thus we conclude that a single skymap will be sufficient to describe the anisotropies in a stochastic background. The multiple capabilities of PyStoch will be useful for estimating and constraining various measures that characterize an anisotropic background.

gr-qc↗

arXiv2019

Stochastic gravitational wave background mapmaking using regularised deconvolution

Obtaining a faithful source intensity distribution map of the sky from noisy data demands incorporating known information of the expected signal, especially when the signal is weak compared to the noise. We introduce a widely used procedure to incorporate these priors through a Bayesian regularisation scheme in the context of map-making of the anisotropic stochastic GW background (SGWB). Specifically, we implement the quadratic form of regularizing function with varying strength of regularization and study its effect on image restoration for different types of the injected source intensity distribution in simulated LIGO data. We find that regularization significantly enhances the quality of reconstruction, especially when the intensity of the source is weak, and dramatically improves the stability of deconvolution. We further study the quality of reconstruction as a function of regularization constant. While in principle this constant is dependent on the data set, we show that the deconvolution process is robust against the choice of the constant, as long as it is chosen from a broad range of values obtained by the method presented here.

gr-qc↗

arXiv2019

Effect of induced seismicity on advanced gravitational wave interferometers

Advanced LIGO and the next generation of ground-based detectors aim to capture many more binary coalescences through improving sensitivity and duty cycle. Earthquakes have always been a limiting factor at low frequency where neither the pendulum suspension nor the active controls provide sufficient isolation to the test mass mirrors. Several control strategies have been proposed to reduce the impact of teleseismic events by switching to a robust configuration with less aggressive feedback. The continental United States has witnessed a huge increase in the number of induced earthquake events primarily associated with hydraulic fracking-related waste water re-injection. Effects from these differ from teleseismic earthquakes primarily because of their depth which is in turn linked to their triggering mechanism. In this paper, we discuss the impact caused due to these low magnitude regional earthquakes and explore ways to minimize the impact of induced seismicity on the detector.

astro-ph.IM↗

arXiv2019

Component separation map-making for stochastic gravitational wave background

Superposition of unresolved sources of gravitational waves (GW) is expected to create a persistent background of stochastic gravitational wave background (SGWB). Different types of astrophysical and cosmological sources are however likely to be present in the background. We present an algorithm for separation of the components with distinct frequency spectral indices into individual sky-maps. To demonstrate the method, we simulate GW signals for different spectral indices and corresponding sky-maps, e.g., point sources like the Virgo cluster and extended ones like the Milky Way Galaxy, and apply our method to recover the injected maps.

gr-qc↗

arXiv2018

Detection and characterization of spin-orbit resonances in the advanced gravitational wave detectors era

In this paper, we test the performance of templates in detection and characterization of Spin-orbit resonant (SOR) binaries. We use precessing SEOBNRv3 waveforms as well as {\it four} numerical relativity (NR) waveforms to model GWs from SOR binaries and filter them through IMRPhenomD, SEOBNRv4 (non-precessing) and IMRPhenomPv2 (precessing) approximants. We find that IMRPhenomD and SEOBNRv4 recover only $\sim70\%$ of injections with fitting factor (FF) higher than 0.97 (or 90\% of injections with ${\rm FF} >0.9$).However, using the sky-maxed statistic, IMRPhenomPv2 performs magnificently better than their non-precessing counterparts with recovering $99\%$ of the injections with FFs higher than 0.97. Interestingly, injections with $Δϕ= 180^{\circ}$ have higher FFs ($Δϕ$ is the angle between the components of the black hole spins in the plane orthogonal to the orbital angular momentum) as compared to their $Δϕ=0^{\circ}$ and generic counterparts. This implies that we will have a slight observation bias towards $Δϕ=180^{\circ}$ SORs while using non-precessing templates for searches. All template approximants are able to recover most of the injected NR waveforms with FFs $>0.95$. For all the injections including NR, the error in estimating chirp mass remains below $<10\%$ with minimum error for $Δϕ= 180^{\circ}$ resonant binaries. The symmetric mass ratio can be estimated with errors below $15\%$. The effective spin parameter $χ_{\rm eff}$ is measured with maximum absolute error of 0.13. The in-plane spin parameter $χ_p$ is mostly underestimated indicating that a precessing signal will be recovered as a relatively less precessing signal. Based on our findings, we conclude that we not only need improvements in waveform models towards precession and non-quadrupole modes but also better search strategies for precessing GW signals.

gr-qc↗

arXiv2018

Very fast stochastic gravitational wave background map making using folded data

A stochastic gravitational-wave background (SGWB) is expected from the superposition of a wide variety of independent and unresolved astrophysical and cosmological sources from different stages in the evolution of the Universe. Radiometric techniques are used to make sky maps of anisotropies in the SGWB by cross-correlating data from pairs of detectors. The conventional searches can be made hundreds of times faster through the folding mechanism introduced recently. Here we present a newly developed algorithm to perform the SGWB searches in a highly efficient way. Taking advantage of the compactness of the folded data we replaced the loops in the pipeline with matrix multiplications. We also incorporated well-known HEALPix pixelization tools for further standardization and optimization. Our Python-based implementation of the algorithm is available as an open source package ${\tt PyStoch}$. Folding and ${\tt PyStoch}$ together has made the radiometer analysis $\textit{a few thousand times}$ faster; it is now possible to make all-sky maps of a stochastic background in just a few minutes on an ordinary laptop. Moreover, ${\tt PyStoch}$ generates a skymap at every frequency bin as an intermediate data product. These techniques have turned SGWB searches very convenient and will make computationally challenging analyses like blind all-sky narrowband search feasible.

gr-qc↗

arXiv2018

Hierarchical search strategy for the efficient detection of gravitational waves from non-precessing coalescing compact binaries with aligned-spins

In the first two years of Gravitational Wave (GW) Astronomy, half a dozen compact binary coalescences (CBCs) have been detected. As the sensitivities and bandwidths of the detectors improve and new detectors join the network, many more sources are expected to be detected. The goal will not only be to find as many sources as possible in the data but to understand the dynamics of the sources much more precisely. Standard searches are currently restricted to a smaller parameter space which assumes aligned spins. Construction of a larger and denser parameter space, and optimising the resultant increase in false alarms, pose a serious computational challenge. We present here a two-stage hierarchical strategy to search for CBCs in data from a network of detectors and demonstrate the computational advantage in real life scenario by introducing it in the standard {\tt PyCBC} pipeline with the usual restricted parameter space. With this strategy, in simulated data containing stationary Gaussian noise, we obtain a computational gain of $\sim 20$ over the flat search. In real data, we expect the computational gain up to a factor of few. This saving in the computational effort will, in turn, allow us to search for precessing binaries. Freeing up computation time for the regular analyses will provide more options to search for sources of different kinds and to fulfil the never-ending urge for extracting more science out of the data with limited resources.

astro-ph.IM↗

arXiv2017

Information Retrieval and Recommendation System for Astronomical Observatories

We present a machine learning based information retrieval system for astronomical observatories that tries to address user defined queries related to an instrument. In the modern instrumentation scenario where heterogeneous systems and talents are simultaneously at work, the ability to supply with the right information helps speeding up the detector maintenance operations. Enhancing the detector uptime leads to increased coincidence observation and improves the likelihood for the detection of astrophysical signals. Besides, such efforts will efficiently disseminate technical knowledge to a wider audience and will help the ongoing efforts to build upcoming detectors like the LIGO-India etc even at the design phase to foresee possible challenges. The proposed method analyses existing documented efforts at the site to intelligently group together related information to a query and to present it on-line to the user. The user in response can further go into interesting links and find already developed solutions or probable ways to address the present situation optimally. A web application that incorporates the above idea has been implemented and tested for LIGO Livingston, LIGO Hanford and Virgo observatories.

astro-ph.IM↗

arXiv2017

Transient Classification in LIGO data using Difference Boosting Neural Network

Detection and classification of transients in data from gravitational wave detectors are crucial for efficient searches for true astrophysical events and identification of noise sources. We present a hybrid method for classification of short duration transients seen in gravitational wave data using both supervised and unsupervised machine learning techniques. To train the classifiers we use the relative wavelet energy and the corresponding entropy obtained by applying one-dimensional wavelet decomposition on the data. The prediction accuracy of the trained classifier on 9 simulated classes of gravitational wave transients and also LIGO's sixth science run hardware injections are reported. Targeted searches for a couple of known classes of non-astrophysical signals in the first observational run of Advanced LIGO data are also presented. The ability to accurately identify transient classes using minimal training samples makes the proposed method a useful tool for LIGO detector characterization as well as searches for short duration gravitational wave signals.

astro-ph.IM↗

arXiv2016

Towards a first design of a Newtonian-noise cancellation system for Advanced LIGO

Newtonian gravitational noise from seismic fields is predicted to be a limiting noise source at low frequency for second generation gravitational-wave detectors. Mitigation of this noise will be achieved by Wiener filtering using arrays of seismometers deployed in the vicinity of all test masses. In this work, we present optimized configurations of seismometer arrays using a variety of simplified models of the seismic field based on seismic observations at LIGO Hanford. The model that best fits the seismic measurements leads to noise reduction limited predominantly by seismometer self-noise. A first simplified design of seismic arrays for Newtonian-noise cancellation at the LIGO sites is presented, which suggests that it will be sufficient to monitor surface displacement inside the buildings.

gr-qc↗

arXiv2016

Statistical isotropy violation in WMAP CMB maps resulting from non-circular beams

Statistical isotropy (SI) of Cosmic Microwave Background (CMB) fluctuations is a key observational test to validate the cosmological principle underlying the standard model of cosmology. While a detection of SI violation would have immense cosmological ramification, it is important to recognise their possible origin in systematic effects of observations. WMAP seven year (WMAP-7) release claimed significant deviation from SI in the bipolar spherical harmonic (BipoSH) coefficients $A_{ll}^{20}$ and $A_{l-2l}^{20}$. Here we present the first explicit reproduction of the measurements reported in WMAP-7, confirming that beam systematics alone can completely account for the measured SI violation. The possibility of such a systematic origin was alluded to in WMAP-7 paper itself and other authors but not as explicitly so as to account for it accurately. We simulate CMB maps using the actual WMAP non-circular beams and scanning strategy. Our estimated BipoSH spectra from these maps match the WMAP-7 results very well. It is also evident that only a very careful and adequately detailed modelling, as carried out here, can conclusively establish that the entire signal arises from non-circular beam effect. This is important since cosmic SI violation signals are expected to be subtle and dismissing a large SI violation signal as observational artefact based on simplistic plausibility arguments run the serious risk of "throwing the baby out with the bathwater".

astro-ph.CO↗

arXiv2015

Estimating SI violation in CMB due to non-circular beam and complex scan in minutes

Mild, unavoidable deviations from circular-symmetry of instrumental beams along with scan strategy can give rise to measurable Statistical Isotropy (SI) violation in Cosmic Microwave Background (CMB) experiments. If not accounted properly, this spurious signal can complicate the extraction of other SI violation signals (if any) in the data. However, estimation of this effect through exact numerical simulation is computationally intensive and time consuming. A generalized analytical formalism not only provides a quick way of estimating this signal, but also gives a detailed understanding connecting the leading beam anisotropy components to a measurable BipoSH characterisation of SI violation. In this paper, we provide an approximate generic analytical method for estimating the SI violation generated due to a non-circular (NC) beam and arbitrary scan strategy, in terms of the Bipolar Spherical Harmonic (BipoSH) spectra. Our analytical method can predict almost all the features introduced by a NC beam in a complex scan and thus reduces the need for extensive numerical simulation worth tens of thousands of CPU hours into minutes long calculations. As an illustrative example, we use WMAP beams and scanning strategy to demonstrate the easability, usability and efficiency of our method. We test all our analytical results against that from exact numerical simulations.

astro-ph.CO↗

arXiv2015

Component Separation of a Isotropic Gravitational Wave Background

A Gravitational Wave Background (GWB) is expected in the universe from the superposition of a large number of unresolved astrophysical sources and phenomena in the early universe. Each component of the background (e.g., from primordial metric perturbations, binary neutron stars, milli-second pulsars etc.) has its own spectral shape. Many ongoing experiments aim to probe GWB at a variety of frequency bands. In the last two decades, using data from ground-based laser interferometric gravitational wave (GW) observatories, upper limits on GWB were placed in the frequency range of ~50-1000 Hz, considering one spectral shape at a time. However, one strong component can significantly enhance the estimated strength of another component. Hence, estimation of the amplitudes of the components with different spectral shapes should be done jointly. Here we propose a method for "component separation" of a statistically isotropic background, that can, for the first time, jointly estimate the amplitudes of many components and place upper limits. The method is rather straightforward and needs negligible amount of computation. It utilises the linear relationship between the measurements and the amplitudes of the actual components, alleviating the need for a sampling based method, e.g., Markov Chain Monte Carlo (MCMC) or matched filtering, which are computationally intensive and cumbersome in a multi-dimensional parameter space. Using this formalism we could also study how many independent components can be separated using a given dataset from a network of current and upcoming ground based interferometric detectors.

astro-ph.CO↗

arXiv2015

All-sky, narrowband, gravitational-wave radiometry with folded data

Gravitational-wave radiometry is a powerful tool by which weak signals with unknown signal morphologies are recovered through a process of cross correlation. Radiometry has been used, e.g., to search for persistent signals from known neutron stars such as Scorpius X-1. In this paper, we demonstrate how a more ambitious search--for persistent signals from unknown neutron stars--can be efficiently carried out using folded data, in which an entire ~year-long observing run is represented as a single sidereal day. The all-sky, narrowband radiometer search described here will provide a computationally tractable means to uncover gravitational-wave signals from unknown, nearby neutron stars in binary systems, which can have modulation depths of ~0.1-2 Hz. It will simultaneously provide a sensitive search algorithm for other persistent, narrowband signals from unexpected sources.

astro-ph.IM↗

arXiv2015

Stochastic Gravitational Wave Background from Exoplanets

Recent exoplanet surveys have predicted a very large population of planetary systems in our galaxy, more than one planet per star on the average, perhaps totalling about two hundred billion. These surveys, based on electro-magnetic observations, are limited to a very small neighbourhood of the solar system and the estimations rely on the observations of only a few thousand planets. On the other hand, orbital motions of planets around stars are expected to emit gravitational waves (GW), which could provide information about the planets not accessible to electro-magnetic astronomy. The cumulative effect of the planets, with periods ranging from few hours to several years, is expected to create a stochastic GW background (SGWB). We compute the characteristic GW strain of this background based on the observed distribution of planet parameters. We also show that the integrated extragalactic background is comparable or less than the galactic background at different frequencies. Our estimate shows that the net background is significantly below the sensitivities of the proposed GW experiments in different frequency bands. However, we notice that the peak of the spectrum, at around $10^{-5}$Hz, is not too far below the proposed space based GW missions. A future space based mission may be able to observe or tightly constrain this signal, which will possibly be the only way to probe the galactic population of exoplanets as a whole.

gr-qc↗

arXiv2015

Fast Gravitational Wave Radiometry using Data Folding

Gravitational Waves (GWs) from the early universe and unresolved astrophysical sources are expected to create a stochastic GW background (SGWB). The GW radiometer algorithm is well suited to probe such a background using data from ground based laser interferometric detectors. Radiometer analysis can be performed in different bases, e.g., isotropic, pixel or spherical harmonic. Each of these analyses possesses a common temporal symmetry which we exploit here to fold the whole dataset for every detector pair, typically a few hundred to a thousand days of data, to only one sidereal day, without any compromise in precision. We develop the algebra and a software pipeline needed to fold data, accounting for the effect of overlapping windows and non-stationary noise. We implement this on LIGO's fifth science run data and validate it by performing a standard anisotropic SGWB search on both folded and unfolded data. Folded data not only leads to orders of magnitude reduction in computation cost, but it results in a conveniently small data volume of few gigabytes, making it possible to perform an actual analysis on a personal computer, as well as easy movement of data. A few important analyses, yet unaccomplished due to computational limitations, will now become feasible. Folded data, being independent of the radiometer basis, will also be useful in reducing processing redundancies in multiple searches and provide a common ground for mutual consistency checks. Most importantly, folded data will allow vast amount of experimentation with existing searches and provide substantial help in developing new strategies to find unknown sources.

gr-qc↗

arXiv2015

Effect of noncircularity of experimental beam on CMB parameter estimation

Measurement of Cosmic Microwave Background (CMB) anisotropies has been playing a lead role in precision cosmology by providing some of the tightest constrains on cosmological models and parameters. However, precision can only be meaningful when all major systematic effects are taken into account. Non-circular beams in CMB experiments can cause large systematic deviation in the angular power spectrum, not only by modifying the measurement at a given multipole, but also introducing coupling between different multipoles through a deterministic bias matrix. Here we add a mechanism for emulating the effect of a full bias matrix to the Planck likelihood code through the parameter estimation code SCoPE. We show that if the angular power spectrum was measured with a non-circular beam, the assumption of circular Gaussian beam or considering only the diagonal part of the bias matrix can lead to huge error in parameter estimation. We demonstrate that, at least for elliptical Gaussian beams, use of scalar beam window functions obtained via Monte Carlo simulations starting from a fiducial spectrum, as implemented in Planck analyses for example, leads to em only few percent of sigma deviation of the best-fit parameters. However, we notice more significant differences in the posterior distributions for some of the parameters, which would in turn lead to incorrect errorbars. These differences can be reduced, so that the errorbars match within few percent, by adding an iterative reanalysis step, where the beam window function would be recomputed using the best-fit spectrum estimated in the first step.

astro-ph.CO↗