Einstein’s general theory of relativity has withstood 100 years of testing and will soon be facing one of its toughest challenges. In a few years we expect to be entering the era of the first direct observations of gravitational waves. These are tiny perturbations of space-time that are generated by accelerating matter and affect the measured distances between two points. Observations of these using the laser interferometers, which are the most sensitive length-measuring devices in the world, will allow us to test models of interactions in the strong field regime of gravity and eventually general relativity itself. I apply the tools of Bayesian inference for the examination of gravitational wave data from the LIGO and Virgo detectors. This is used for signal detection and estimation of the source parameters. I quantify the ability of a network of ground-based detectors to localise a source position on the sky for electromagnetic follow-up. Bayesian criteria are also applied to separating real signals from glitches in the detectors. These same tools and lessons can also be applied to the type of data expected from planned space-based detectors. Using simulations from the Mock LISA Data Challenges, I analyse our ability to detect and characterise both burst and continuous signals. The two seemingly different signal types will be overlapping and confused with one another for a space-based detector; my analysis shows that we will be able to separate and identify many signals present. Data sets and astrophysical models are continuously increasing in complexity. This will create an additional computational burden for performing Bayesian inference and other types of data analysis. I investigate the application of the MOPED algorithm for faster parameter estimation and data compression. I find that its shortcomings make it a less favourable candidate for further implementation. The framework of an artificial neural network is a simple model for the structure of a brain which can “learn” functional relationships between sets of inputs and outputs. I describe an algorithm developed for the training of feed-forward networks on pre-calculated data sets. The trained networks can then be used for fast prediction of outputs for new sets of inputs. After demonstrating capabilities on toy data sets, I apply the ability of the network to classifying handwritten digits from the MNIST database and measuring ellipticities of galaxies in the Mapping Dark Matter challenge. The power of neural networks for learning and rapid prediction is also useful in Bayesian inference where the likelihood function is computationally expensive. The new BAMBI algorithm is detailed, in which our network training algorithm is combined with the nested sampling algorithm MULTINEST to provide rapid Bayesian inference. Using samples from the normal inference, a network is trained on the likelihood function and eventually used in its place. This is able to provide significant increase in the speed of Bayesian inference while returning identical results. The trained networks can then be used for extremely rapid follow-up analyses with different priors, obtaining orders of magnitude of speed increase. Learning how to apply the tools of Bayesian inference for the optimal recovery of gravitational wave signals will provide the most scientific information when the first detections are made. Complementary to this, the improvement of our analysis algorithms to provide the best results in less time will make analysis of larger and more complicated models and data sets practical.

We have performed a blind search for a gamma-ray transient of arbitrary duration and energy spectrum around the time of the LIGO gravitational-wave event GW150914 with the six-spacecraft interplanetary network (IPN). Four gamma-ray bursts were detected between 30 hr prior to the event and 6.1 hr after it, but none could convincingly be associated with GW150914. No other transients were detected down to limiting 15-150 keV fluences of roughly 5 x10-(8) -5 x 10(-7) erg cm(-2). We discuss the search strategies and temporal coverage of the IPN on the day of the event and compare the spatial coverage to the region where GW150914 originated. We also report the negative result of a targeted search for the Fermi-GBM event reported in conjunction with GW150914.

In his theory of General Relativity, Einstein describes gravity as a geometric property of spacetime, which deforms in the presence of mass and energy. The accelerated motion of masses produces deformations, which propagate outward from their source at the speed of light. We refer to these radiated deformations as gravitational waves. Over the past several decades, the goal of the Laser Interferometer Gravitational-wave Observatory (LIGO) has been the search for direct evidence of gravitational waves from astrophysical sources, using ground based laser interferometers. As LIGO moves into its Advanced era (aLIGO), the direct detection of gravitational waves is inevitable. With the technology at hand, it is imperative that we have the tools to analyze the detector signal and examine the interesting astrophysical properties of the source. Some of the main targets of this search are coalescing compact binaries. In this thesis, I describe and evaluate bhextractor, a data analysis algorithm that uses Principal Component Analysis (PCA) to identify the main features of a set of gravitational waveforms produced by the coalescence of two black holes. Binary Black Hole (BBH) systems are expected to be among the most common sources of gravitational waves in the sensitivity band of aLIGO. However, the gravitational waveforms emitted by BBH systems are not well modeled and require computationally expensive Numerical Relativity (NR) simulations. bhextractor uses PCA to decompose a catalog of available NR waveforms into a set of orthogonal Principal Components (PCs), which efficiently select the major common features of the waveforms in the catalog and represent a portion of the BBH parameter space. From these PCs, we can reconstruct any waveform in the catalog, and construct new waveforms with similar properties. Using Bayesian analysis and Nested Sampling, one can use bhextractor to classify an arbitrary BBH waveform into one of the available catalogs and estimate the parameters of the gravitational wave source.

京都大学 / 新制・課程博士 / 博士(理学) / 甲第23361号 / 理博第4732号 / 新制||理||1679(附属図書館) / 京都大学大学院理学研究科物理学・宇宙物理学専攻 / (主査)教授 田中 貴浩, 准教授 久徳 浩太郎, 教授 萩野 浩一 / 学位規則第4条第1項該当 / Doctor of Science / Kyoto University / DFAM

京都大学 / 0048 / 新制・課程博士 / 博士(理学) / 甲第20900号 / 理博第4352号 / 新制||理||1625(附属図書館) / 京都大学大学院理学研究科物理学・宇宙物理学専攻 / (主査)教授 田中 貴浩, 教授 向山 信治, 教授 川合 光 / 学位規則第4条第1項該当 / Doctor of Science / Kyoto University / DFAM

No description available.

Since the first observation of a gravitational-wave (GW) in 2015, the LIGO and Virgo detectors reported tens of astrophysical signals interpreted as mergers of compact objects. These observations provide invaluable tests of the General Relativity and open a new era of astronomy, unveiling compact objects’ nature. The focus of the thesis is the detection and the characterization of GW transients with minimal assumptions on the GW sources. To identify astrophysical signals embedded in detector noise, there are two main approaches: template-based and unmodelled searches. The firsts look for GW signals with a time-frequency evolution consistent to the waveform models contained in extensive template banks. Instead, unmodelled or burst searches do not assume a waveform model, but look for excess of power that is coherent on multiple GW detectors. Burst search are fundamental to observe GWs from various astrophysical sources. Unmodelled searches observe GWs originated from the coalescence of compact binaries, and might observe GWs that are expected by other sources such as supernovae, isolated neutron stars, and cosmic strings. Burst searches also provide the reconstruction of the GW waveform with minimal assumptions, and are able to identify discrepancies between theoretical models and measured data, which may reveal new physics. A well-known software for burst searches is Coherent WaveBurst (cWB). cWB identifies excess of power with respect to the detector noise that are coherent in the GW detectors network. Within this framework, the thesis presents three author’s original contributions to this field. The first is the search sensitivity of three-detectors network in burst searches. Having more detectors participating in the GW observations generally improves the source localization and the characterization of the GW signals. The capability of burst searches to distinguish between potential signals and transient noise depends on the orientation of the detectors and on their relative sensitivities. In literature, the cWB search sensitivity of the three-detectors network composed of the LIGO and Virgo detectors (HLV) is lower than the one achieved using only LIGO detectors (HL). cWB uses likelihood regulators to force the reconstruction of the GW component observed by the LIGO aligned detectors. These regulators successfully reduce the false alarm rate of the HL coherent analysis, but to make full use of a third, not-aligned detector, they should be relaxed. The fifth chapter investigates the impact of the likelihood regulators in cWB for HLV network, first in a simplified case assuming Gaussian noise only, and then in the data from the third LIGO-Virgo-KAGRA observing run. Thanks to latest cWB enhancements and relaxed likelihood regulators, we show that the HLV network reduces significantly the gap w.r.t. HL, having a higher sensitivity for several waveforms tested on average over the sky directions. Moreover, we investigate the use of the HLV network to test the consistency between cWB unmodelled signal reconstruction and the GW waveform models. The second original contribution is the development of an autoencoder neural network integrated into GW burst searches to improve the rejection of noise transients GW data contains short-duration disturbances, called glitches, which can mimic astrophysical signals. Mitigation of glitches is particularly difficult for unmodelled algorithm, such as cWB, that do not use GW waveform models to filter the data, but are sensitive to the widest possible range of morphologies. Noise mitigation is a long-term effort in cWB, which led to the introduction of specific estimators and a machine-learning based signal-noise classification algorithm. The sixth chapter presents an autoencoder neural network, integrated into cWB, that learns transient noise morphologies from GW time-series and it improves their rejection. An autoencoder is an unsupervised learning neural network that compresses the input data into a lower dimensional space, called latent space, and then re-constructs an output with the original dimensions. Here, the autoencoder is trained on time-series belonging to a single glitch family, known as blip, and the network learns that specific morphology. The autoencoder improves cWB discrimination between blip-like glitches and potential GW signals, reducing the background trigger at low frequencies. We inject in the LIGO detectors’ data from the third Advanced LIGO-Virgo observing run a wide range of simulated signals, and we evaluate the cWB search sensitivity including the autoencoder output in the cWB ranking statistics. At a false alarm rate of one event per 50 years, the sensitivity volume increases up to 30% for signal morphologies similar to blip glitches. Finally, the thesis presents the search for hyperbolic encounters between compact objects in the data from the third LIGO-Virgo-KAGRA observing run. As GW detectors sensitivity increases, new astrophysical sources could emerge. Close hyperbolic encounters (HE) are one such source class: scattering of stellar mass compact objects is expected to manifest as GW burst signals in the frequency band of current detectors. The seventh chapter presents the search for GWs from HE in the data from the second-half of the third observing run using cWB. No significant event has been identified in addition to known detections of GW events. We inject third Post-Newtonian order accurate HE waveforms with component masses between [2,100]M ⊙ . For the first time, we report the sensitivity volume achieved for such sources, i.e. the portion of the Universe in which the proposed analysis would have detected a HE signal with a certain significance, if any. The sensitivity volume peaks at 3.9±1.4×10 5 Mpc3 year for compact objects with masses between [20, 40] M ⊙, corresponding to a rate density upper limit of 0.589±0.094 ×10 −5 Mpc −3 year −1. Moreover, the sensitive volume prospects for the next observing runs of current detectors are discussed. All the result shown are based on the latest publicly available data from the third observing run of the LIGO-Virgo-KAGRA collaboration.

Gravitational waves are truly transparent to matter in the Universe and carry the information of the very early epoch. We show that the energy density spectrum of the primordial gravitational waves has characteristic features due to the successive changes in the relativistic degrees of freedom during the radiation era. Our calculations are solely based on the standard model of cosmology and particle physics, and therefore these features must exist. Our calculations significantly improve the previous ones which ignored these effects and predicted a smooth, featureless spectrum. Going back in time to the beginning of the radiation era, reheating of the Universe must have taken place after inflation for primordial nucleosynthesis to begin. We show that reheating occurs spontaneously in a broad class of inflation models with [scientific symbols] gravity (Ø is inflaton). The model does not require explicit couplings between Ø and bosonic or fermionic matter fields. The couplings arise spontaneously when Ø settles in the vacuum expectation value (vev) and oscillates. This mechanism allows inflaton quanta to decay into any fields which are not conformally invariant in [scientific symbols] gravity theories. Applying the above method, we study implications of the large-N species solution to the hierarchy problem, proposed by G. Dvali, for reheating after inflation. We show that, in this scenario, the decay rates of inflaton fields through gravitational decay channels are enhanced by a factor of N, and thus they decay into N species of the quantum fields very efficiently. Without violating energy conservation, cosmological consideration places non-trivial constraints on Dvali's solution to the hierarchy problem. Going back in time still further, we study the period just before the beginning of reheating, the era of coherent oscillation of scalar fields. We show that non-Gaussian primordial curvature perturbations appear temporarily in the coherent oscillation phase after multi-field inflation. We directly solve the evolution equation of non-Gaussianity on super-horizon scales caused by the non-linear influence of entropy perturbations on the curvature perturbations during this phase. We show that our approach precisely matches with the so-called "separate universe approach" or "δN formalism" by studying a simple quadratic two-field potential. / text

The collapse of a stellar core is expected to produce gravitational waves (GWs), neutrinos, and in most cases a luminous supernova. Sometimes, however, the optical event could be significantly less luminous than a supernova and a direct collapse to a black hole, where the star just disappears, is possible. The GW event GW150914 was detected by the LIGO Virgo Collaboration via a burst analysis that gave localization contours enclosing the Large Magellanic Cloud (LMC). Shortly thereafter, we used DECam to observe 102 deg(2) of the localization area, including 38 deg(2) on the LMC for a missing supergiant search. We construct a complete catalog of LMC luminous red supergiants, the best candidates to undergo invisible core collapse, and collected catalogs of other candidates: less luminous red supergiants, yellow supergiants, blue supergiants, luminous blue variable stars, and Wolf-Rayet stars. Of the objects in the imaging region, all are recovered in the images. The timescale for stellar disappearance is set by the free-fall time, which is a function of the stellar radius. Our observations at 4 and 13 days after the event result in a search sensitive to objects of up to about 200 solar radii. We conclude that it is unlikely that GW150914 was caused by the core collapse of a relatively compact supergiant in the LMC, consistent with the LIGO Collaboration analyses of the gravitational waveform as best interpreted as a high mass binary black hole merger. We discuss how to generalize this search for future very nearby core-collapse candidates.

We report the results of a deep search for an optical counterpart to the gravitational wave (GW) event GW150914, the first trigger from the Advanced LIGO GW detectors. We used the Dark Energy Camera (DECam) to image a 102 deg(2) area, corresponding to 38% of the initial trigger high-probability sky region and to 11% of the revised high-probability region. We observed in the i and z bands at 4-5, 7, and 24 days after the trigger. The median 5 sigma point-source limiting magnitudes of our search images are i = 22.5 and z = 21.8 mag. We processed the images through a difference-imaging pipeline using templates from pre-existing Dark Energy Survey data and publicly available DECam data. Due to missing template observations and other losses, our effective search area subtends 40 deg(2), corresponding to a 12% total probability in the initial map and 3% in the final map. In this area, we search for objects that decline significantly between days 4-5 and day 7, and are undetectable by day 24, finding none to typical magnitude limits of i = 21.5, 21.1, 20.1 for object colors (i - z) = 1, 0, - 1, respectively. Our search demonstrates the feasibility of a dedicated search program with DECam and bodes well for future research in this emerging field.