Support of LIGO Data Analysis Activities at the University of Texas at Brownsville

爱因斯坦发现的广义相对论告诉我们，我们日常所熟悉的引力其实是某种更奇怪的东西的表现：物质对时空几何形状的弯曲。该理论的关键预测包括宇宙膨胀和黑洞的存在，其中包括引力波（GW）的存在：在大质量的快速运动引起的时空几何中以光速运动的涟漪。虽然引力波对致密双星系统的间接影响已经得到了很好的验证，但直接探测地球上的引力波仍然是一个突出的挑战。实现这种能力的科学回报将是巨大的——从探测爆炸恒星的极端动力学到收集关于宇宙大爆炸时刻的状态的信息。这扇探索宇宙的新窗口耗费了数十年的实验和技术发展，突破了物理科学不同领域的界限。2015年将标志着引力波物理学备受期待的分水岭时刻：两个先进的激光干涉引力波天文台（aLIGO）探测器将开始它们的初始数据采集运行，随后是欧洲先进的处女座引力波天文台的调试。aLIGO探测器的灵敏度将提高到第一代探测器的10倍左右，为观测GW源打开了比以前大1000倍的空间体积。随着仪器仪表的这些巨大进步，从引力波物理学研究的一开始，人们就知道统计数据分析技术的范围也必须进一步推进，以便能够探测到嵌入在噪声中的微弱和罕见的信号。正是在这一领域，本基金资助的研究将作出重大贡献。通过支持研究生，这项资助将有助于发展该领域的研究人员群体。由于德克萨斯大学布朗斯维尔分校（UTB）是一所西班牙裔服务机构，这些研究活动将使传统上在STEM领域代表性不足的学生接触到前沿科学。UTB正在进行的主要教育和推广活动将利用这些活动来提高高中生对LIGO等令人兴奋的项目的认识。本基金支持以下主要数据分析领域的研究项目。(i)将拟订和执行一种算法，使其能够在最少的事先假设下，构造随机引力波各向异性背景的天空图。使用最初在基于脉冲星时间的引力波搜索背景下开发的方法，新方法将极大地扩展现有依赖模型的搜索的能力。标准频率随机搜索的贝叶斯替代方案也将被开发，它将在检测器表征和噪声估计的背景下利用现有的贝叶斯推理代码。（ii）将实施一种基于平滑正则化的相干网络分析方法，该方法将显著改善具有非紧凑时频（TF）特征的突发引力波信号的检测和表征。这种信号预计会出现在各种天体物理场景中，但已知会对现有方法构成重大挑战。（iii）从声学信号处理中发展的谐波再生降噪（HRNR）技术衍生出的一种新方法，将与现有的网络分析管道相结合，以提高其对核后反弹相位超新星信号的灵敏度。通过改进信号波形和天空定位估计，这将导致更好的多信使跟踪这种罕见但戏剧性的事件。数据中探测器噪声和非引力波信号（“小故障”）的特征对于提高对真正引力波信号的探测置信度是必要的。在这种情况下，将使用自组织图（SOM）方法进行仪器故障分类，并对与aLIGO相关的故障特征进行编目。一种独立的方法将使用BayesWave贝叶斯推理管道来识别和研究故障信号。

The General Theory of Relativity discovered by Einstein tells us that the familiar, everyday force of gravity is a manifestation of something much stranger: the bending of the geometry of space-time by matter. Among the key predictions of the theory, which includes the expanding Universe and the existence of black holes, is the existence of gravitational waves (GW): ripples moving at the speed of light in the geometry of space-time caused by the fast motion of large masses. Although well tested in terms of their indirect effects on binary systems of compact stars, the direct detection of gravitational waves incident on Earth poses an outstanding challenge. The scientific rewards from achieving this ability would be enormous - ranging from probing the extreme dynamics of exploding stars to gleaning information about the state of the Universe almost at the moment of the Big Bang itself. The effort to enable this new window on the universe has occupied several decades of experimental and technological developments that have pushed the boundaries across diverse fields in the physical sciences. The year 2015 will mark a highly-anticipated watershed moment for gravitational-wave physics: The two advanced Laser Interferomenter Gravitational Wave Observatory (aLIGO) detectors will start their initial data taking runs, followed by the commissioning of the advanced Virgo gravitational wave observatory in Europe. The sensitivity of the aLIGO detector will be ramped up to become about ten times better than that of the first-generation detectors, opening up a spatial volume for observing GW sources that will be 1000 times larger than before. Along with these tremendous advances in instrumentation, it has been known from the very inception of the gravitational wave physics effort that the envelope of statistical data analysis techniques must also be pushed further to enable the detection of weak and rare signals embedded in noise. It is in this area that the research funded by this grant will make significant contributions. By supporting graduate students, this grant will help to grow the community of researchers in this field. Since the University of Texas at Brownsville (UTB) is an Hispanic serving institution, the research activities will expose students who are traditionally under-represented in STEM areas to forefront science. Ongoing major education and outreach activities at UTB will leverage these activities to create awareness among high-school students about exciting projects such as LIGO.This grant supports research projects in the following major data analysis areas. (i) An algorithm will be developed and implemented that allows the construction of sky maps of an anisotropic background of stochastic gravitational waves with minimal prior assumptions. Using methods originally developed in the context of pulsar-timing-based gravitational-wave searches, the new approach will vastly extend the capabilities of existing model-dependent searches. A Bayesian alternative to the standard frequentist stochastic search will also be developed that will leverage existing code for Bayesian inference in the context of detector characterization and noise estimation. (ii) A smoothness-regularization-based coherent network analysis method will be implemented that significantly improves the detection and characterization of burst gravitational wave signals having non-compact time-frequency (TF) signatures. Such signals are expected to arise in a variety of astrophysical scenarios but are known to pose a significant challenge to existing methods. (iii) A new method, derived from the Harmonic Regeneration Noise Reduction (HRNR) technique developed in acoustical signal processing, will be integrated with existing network analysis pipelines to boost their sensitivity to post-core-bounce-phase supernova signals. By improving both signal waveform and sky location estimation, this will lead to better multi-messenger follow up of such a rare but dramatic event. (iv) The characterization of detector noise and non-gravitational wave signals ("glitches") in the data is necessary for improving detection confidence for genuine gravitational-wave signals. In this context, the Self Organizing Map (SOM) method will be used for instrumental glitch classification, and glitch characteristics relevant to aLIGO will be catalogued. An independent approach will use the BayesWave Bayesian inference pipeline to identify and study glitch signals.