Bulletin of the American Physical Society
APS April Meeting 2023
Volume 68, Number 6
Minneapolis, Minnesota (Apr 15-18)
Virtual (Apr 24-26); Time Zone: Central Time
Session F09: Gravitational Wave Data Analysis: Machine Learning Methods and Black Hole Spin Inference |
Hide Abstracts |
Sponsoring Units: DGRAV GDS Chair: Katerina Chatziioannou, Caltech Room: Conrad B/C - 2nd Floor |
Sunday, April 16, 2023 8:30AM - 8:42AM |
F09.00001: Bridging the machine learning deployment gap in gravitational wave physics Alec M Gunny, Ethan J Marx, William Benoit, Deep Chatterjee, Michael W Coughlin, Erik Katsavounidis, Muhammed Saleem, Eric Moreno, Philip C Harris, Dylan S Rankin, Ryan J Raikman Deep learning algorithms have achieved state-of-the-art performance across a wide array of settings in recent years, and with this success has come an abundance of papers applying these methods to gravitational wave physics problems. Despite this, there are still comparatively few deep learning models planned for online deployment during the O4 data collection run of the LIGO-Virgo-KAGRA collaboration. One reason for this disparity is a lack of standardized software tooling adequate to quickly implement and iterate upon novel ideas and validate them on sufficiently large volumes of data to achieve confidence in production performance. In this work, we describe a suite of libraries intended to bridge this gap, ml4gw and hermes, and outline how their use in two specific applications has led to increases in efficiency and model robustness. |
Sunday, April 16, 2023 8:42AM - 8:54AM |
F09.00002: Detecting and Denoising Gravitational Waves from Neutron Stars using Deep Learning Chinthak Murali, David Lumley We present the second part of our Deep Learning based denoising scheme to detect and denoise astrophysical gravitational waves. Here we extend our analysis to include gravitational waves from coalescing Binary Neutron Stars (BNS) and Neutron Star- Black Hole (NSBH) binaries in addition to the Binary Black Holes (BBH) presented previously. We employ similar strategy to perform the denoising on the detector data, with slightly different Neural-Net architecture. We use a convolutional neural network, designed in the auto-encoder configuration that can detect and denoise astrophysical gravitational waves from compact binary coalescence, orders of magnitude faster than the conventional matched-filtering based detection that is currently employed at advanced LIGO (aLIGO) and LVK collaboration at large. The Neural-Net architecture is such that it learns from the sparse representation of data in the time-frequency domain and constructs a non-linear mapping function that maps this representation into two separate masks for signal and noise, facilitating the separation of the two, from raw data. This approach is the first of its kind to apply machine learning based gravitational wave detection/denoising in the 2D representation of gravitational wave data. We applied our formalism to the first gravitational wave from binary neutron star, GW170817, successfully recovering the signal at all three phases of coalescence at both detectors with very low False Alarm Probability (FAP) . The recovered signal of this very high SNR detection differs slightly from the template waveform due of the presense of strong precession signature present in the denoised waveform. This method can also interpolate and extrapolate between modeled templates and explore gravitational waves that are unmodeled and hence not present in the template bank of signals used in the matched-filtering detection pipelines. Faster and efficient detection schemes, such as this method, will be instrumental as ground based detectors reach their design sensitivity, likely to result in several hundreds of potential detections in a few months of observing runs. |
Sunday, April 16, 2023 8:54AM - 9:06AM |
F09.00003: New search pipeline for gravitational waves with higher-order harmonics Digvijay S Wadekar, Tejaswi Venumadhav, Matias Zaldarriaga, Javier Roulet, Barak Zackay, Seth Olsen, Jonathan Mushkin, Ajit Mehta
|
Sunday, April 16, 2023 9:06AM - 9:18AM |
F09.00004: Quasi-Anomalous Gravitational-Wave Detection with Recurrent Autoencoders Ryan J Raikman, Eric Moreno, Erik Katsavounidis, Philip C Harris, Ethan J Marx, William Benoit, Ekaterina Govorkova, Deep Chaterjee, Michael W Coughlin, Muhamed Saleem, Dylan S Rankin Detection of gravitational wave (GW) signals in laser interferometers relies on having well modeled templates of the GW emission. We present a method of anomaly detection techniques based on deep recurrent autoencoders to the enhance the search region to potential, unmodelled transients. We use a semi-supervised strategy dubbed Quasi Anomalous Knowledge (QUAK) which provides a weak distinction between classes at training time. While the semi-supervised nature of the problem comes with a cost in terms of accuracy as compared to supervised techniques, there is a qualitative advantage in generalizing experimental sensitivity beyond pre-computed signal templates. We construct a low-dimensional embedded space using the QUAK method which captures the physical signatures of distinct signals on each axis of the space. By introducing alternative signal priors that capture some of the salient features of gravitational-wave signatures, we allow for the recovery of sensitivity even when an unmodelled anomaly is encountered. We show that regions of the QUAK space can identify binaries, detector glitches and also search a variety of hypothesized astrophysical sources that may emit GWs in the LIGO frequency band, including core-collapse supernovae and other stochastic sources. |
Sunday, April 16, 2023 9:18AM - 9:30AM |
F09.00005: Detecting Binary Black Hole Mergers with Effective Machine Learning Infrastructure William Benoit, Alec M Gunny, Ethan J Marx, Deep Chatterjee, Rafia Omer, Michael W Coughlin, Erik Katsavounidis, Muhammed Saleem, Eric Moreno, Dylan S Rankin, Philip C Harris, Ryan J Raikman Deep learning models have quickly become a popular alternative to traditional matched filtering analyses for the identification of gravitational wave signals. The reduced computational cost and the potential for real-time, higher confidence detections have made these techniques an attractive avenue to explore; however, work remains in developing a network that can be effectively deployed on live data during a data collection run of the LIGO-Virgo-KAGRA detectors. We present here the preliminary results of a model for identifying binary black hole mergers, BBHNet, which has been built using libraries that address this gap, hermes and ml4gw. We demonstrate that our model is capable of event identification, and show that our design choices allow for rapid iteration and effective analysis of model performance. |
Sunday, April 16, 2023 9:30AM - 9:42AM |
F09.00006: Spin it as you like: the (lack of a) measurement of the spin tilt distribution with LIGO-Virgo-KAGRA binary black holes Salvatore Vitale, Andrea S Biscoveanu, Colm Talbot The growing set of gravitational-wave sources is being used to measure the properties of the underlying astrophysical populations of compact objects, black holes and neutron stars. Most of the detected systems are black hole binaries. While much has been learned about black holes by analyzing the latest LIGO-Virgo-KAGRA (LVK) catalog, GWTC-3, a measurement of the astrophysical distribution of the black hole spin orientations remains elusive. This is usually probed by measuring the cosine of the tilt angle ($cos{ au}$) between each black hole spin and the orbital angular momentum, $cos{ au}=+1$ being perfect alignment. The LVK has modeled the $cos{ au}$ distribution as a mixture of an isotropic component and a Gaussian component with mean fixed at $+1$ and width measured from the data. In this paper, we want to verify if the data require the existence of such a peak at $cos{ au}=+1$. We use various models for the astrophysical tilt distribution and find that a) Augmenting the LVK model such that the mean $mu$ of the Gaussian is not fixed at $+1$ returns results that strongly depend on priors. If we allow $mu>+1$ then the resulting astrophysical $cos{ au}$ distribution peaks at $+1$ and looks linear, rather than Gaussian. If we constrain $-1leq mu leq +1$ the Gaussian component peaks at $mu=0.47^{+0.47}_{-1.04}$ (median and 90\% symmetric credible interval). Two other 2-component mixture models yield $cos{ au}$ distributions that either have a broad peak centered at $0.20^{+0.21}_{-0.18}$ or a plateau that spans the range $[-0.5, +1]$, without a clear peak at $+1$. b) All of the models we considered agree on the fact that there is no excess of black hole tilts at around $-1$. c) While yielding quite different posteriors, the models considered in this work have Bayesian evidences that are the same within error bars. |
Sunday, April 16, 2023 9:42AM - 9:54AM |
F09.00007: How can we measure spin precession for heavy binary black holes using gravitational waves? Simona J Miller, Maximiliano Isi, Katerina Chatziioannou, Vijay Varma The spins of black holes (BHs) in binary black hole (BBH) systems offer a unique probe of physics on multiple scales, from stellar interiors to the astrophysical environments in which compact binaries form. Despite their astrophysical importance, spin magnitudes and tilt angles for BBHs remain poorly constrained using gravitational-wave (GW) data from the Laser Interferometer GW Observatory (LIGO). The components of spin lying in the orbital plane, causing precession, are particularly hard to measure due to their weak imprint on GW signals and ability to mimic other physical effects. As a spurious measurement of precession could arise from a non-astrophysical source (i.e. small fluctuations in detector noise, data quality issues, waveform systematics, etc.), it is imperative that any claims of detected precession are supported via a robust understanding of where in waveforms and for what types of systems precession is measurable. Using results from time-domain parameter estimation with state-of-the-art numerical relativity surrogate waveforms, I will discuss the roles of the inspiral or merger phases of coalescence in measuring precession for high-mass BBHs. I present LIGO's BBH event GW190521 as a case study, and explore simulated signals of related systems. |
Follow Us |
Engage
Become an APS Member |
My APS
Renew Membership |
Information for |
About APSThe American Physical Society (APS) is a non-profit membership organization working to advance the knowledge of physics. |
© 2024 American Physical Society
| All rights reserved | Terms of Use
| Contact Us
Headquarters
1 Physics Ellipse, College Park, MD 20740-3844
(301) 209-3200
Editorial Office
100 Motor Pkwy, Suite 110, Hauppauge, NY 11788
(631) 591-4000
Office of Public Affairs
529 14th St NW, Suite 1050, Washington, D.C. 20045-2001
(202) 662-8700