-
Convolutional Neural Networks for signal detection in real LIGO data
Authors:
Ondřej Zelenka,
Bernd Brügmann,
Frank Ohme
Abstract:
Searching the data of gravitational-wave detectors for signals from compact binary mergers is a computationally demanding task. Recently, machine learning algorithms have been proposed to address current and future challenges. However, the results of these publications often differ greatly due to differing choices in the evaluation procedure. The Machine Learning Gravitational-Wave Search Challeng…
▽ More
Searching the data of gravitational-wave detectors for signals from compact binary mergers is a computationally demanding task. Recently, machine learning algorithms have been proposed to address current and future challenges. However, the results of these publications often differ greatly due to differing choices in the evaluation procedure. The Machine Learning Gravitational-Wave Search Challenge was organized to resolve these issues and produce a unified framework for machine-learning search evaluation. Six teams submitted contributions, four of which are based on machine learning methods and two are state-of-the-art production analyses. This paper describes the submission from the team TPI FSU Jena and its updated variant. We also apply our algorithm to real O3b data and recover the relevant events of the GWTC-3 catalog.
△ Less
Submitted 12 February, 2024;
originally announced February 2024.
-
MLGWSC-1: The first Machine Learning Gravitational-Wave Search Mock Data Challenge
Authors:
Marlin B. Schäfer,
Ondřej Zelenka,
Alexander H. Nitz,
He Wang,
Shichao Wu,
Zong-Kuan Guo,
Zhoujian Cao,
Zhixiang Ren,
Paraskevi Nousi,
Nikolaos Stergioulas,
Panagiotis Iosif,
Alexandra E. Koloniari,
Anastasios Tefas,
Nikolaos Passalis,
Francesco Salemi,
Gabriele Vedovato,
Sergey Klimenko,
Tanmaya Mishra,
Bernd Brügmann,
Elena Cuoco,
E. A. Huerta,
Chris Messenger,
Frank Ohme
Abstract:
We present the results of the first Machine Learning Gravitational-Wave Search Mock Data Challenge (MLGWSC-1). For this challenge, participating groups had to identify gravitational-wave signals from binary black hole mergers of increasing complexity and duration embedded in progressively more realistic noise. The final of the 4 provided datasets contained real noise from the O3a observing run and…
▽ More
We present the results of the first Machine Learning Gravitational-Wave Search Mock Data Challenge (MLGWSC-1). For this challenge, participating groups had to identify gravitational-wave signals from binary black hole mergers of increasing complexity and duration embedded in progressively more realistic noise. The final of the 4 provided datasets contained real noise from the O3a observing run and signals up to a duration of 20 seconds with the inclusion of precession effects and higher order modes. We present the average sensitivity distance and runtime for the 6 entered algorithms derived from 1 month of test data unknown to the participants prior to submission. Of these, 4 are machine learning algorithms. We find that the best machine learning based algorithms are able to achieve up to 95% of the sensitive distance of matched-filtering based production analyses for simulated Gaussian noise at a false-alarm rate (FAR) of one per month. In contrast, for real noise, the leading machine learning search achieved 70%. For higher FARs the differences in sensitive distance shrink to the point where select machine learning submissions outperform traditional search algorithms at FARs $\geq 200$ per month on some datasets. Our results show that current machine learning search algorithms may already be sensitive enough in limited parameter regions to be useful for some production settings. To improve the state-of-the-art, machine learning algorithms need to reduce the false-alarm rates at which they are capable of detecting signals and extend their validity to regions of parameter space where modeled searches are computationally expensive to run. Based on our findings we compile a list of research areas that we believe are the most important to elevate machine learning searches to an invaluable tool in gravitational-wave signal detection.
△ Less
Submitted 22 September, 2022;
originally announced September 2022.
-
Training Strategies for Deep Learning Gravitational-Wave Searches
Authors:
Marlin B. Schäfer,
Ondřej Zelenka,
Alexander H. Nitz,
Frank Ohme,
Bernd Brügmann
Abstract:
Compact binary systems emit gravitational radiation which is potentially detectable by current Earth bound detectors. Extracting these signals from the instruments' background noise is a complex problem and the computational cost of most current searches depends on the complexity of the source model. Deep learning may be capable of finding signals where current algorithms hit computational limits.…
▽ More
Compact binary systems emit gravitational radiation which is potentially detectable by current Earth bound detectors. Extracting these signals from the instruments' background noise is a complex problem and the computational cost of most current searches depends on the complexity of the source model. Deep learning may be capable of finding signals where current algorithms hit computational limits. Here we restrict our analysis to signals from non-spinning binary black holes and systematically test different strategies by which training data is presented to the networks. To assess the impact of the training strategies, we re-analyze the first published networks and directly compare them to an equivalent matched-filter search. We find that the deep learning algorithms can generalize low signal-to-noise ratio (SNR) signals to high SNR ones but not vice versa. As such, it is not beneficial to provide high SNR signals during training, and fastest convergence is achieved when low SNR samples are provided early on. During testing we found that the networks are sometimes unable to recover any signals when a false alarm probability $<10^{-3}$ is required. We resolve this restriction by applying a modification we call unbounded Softmax replacement (USR) after training. With this alteration we find that the machine learning search retains $\geq 91.5\%$ of the sensitivity of the matched-filter search down to a false-alarm rate of 1 per month.
△ Less
Submitted 10 February, 2022; v1 submitted 7 June, 2021;
originally announced June 2021.
-
Growth of resonances and chaos for a spinning test particle in the Schwarzschild background
Authors:
Ondřej Zelenka,
Georgios Lukes-Gerakopoulos,
Vojtěch Witzany,
Ondřej Kopáček
Abstract:
Inspirals of stellar mass compact objects into supermassive black holes are known as extreme mass ratio inspirals. In the simplest approximation, the motion of the compact object is modeled as a geodesic in the space-time of the massive black hole with the orbit decaying due to radiated energy and angular momentum, thus yielding a highly regular inspiral. However, once the spin of the secondary co…
▽ More
Inspirals of stellar mass compact objects into supermassive black holes are known as extreme mass ratio inspirals. In the simplest approximation, the motion of the compact object is modeled as a geodesic in the space-time of the massive black hole with the orbit decaying due to radiated energy and angular momentum, thus yielding a highly regular inspiral. However, once the spin of the secondary compact body is taken into account, integrability is broken and prolonged resonances along with chaotic motion appear.
We numerically integrate the motion of a spinning test body in the field of a non-spinning black hole and analyse it using various methods. We show for the first time that resonances and chaos can be found even for astrophysical values of spin. On the other hand, we devise a method to analyse the growth of the resonances, and we conclude that the resonances we observe are only caused by terms quadratic in spin and will generally stay very small in the small-mass-ratio limit. Last but not least, we compute gravitational waveforms by solving numerically the Teukolsky equations in the time-domain and establish that they carry information on the motion's dynamics. In particular, we show that the time series of the gravitational wave strain can be used to discern regular from chaotic motion of the source.
△ Less
Submitted 1 November, 2019;
originally announced November 2019.
-
Recurrence analysis of spinning particles in the Schwarzschild background
Authors:
Ondřej Zelenka,
Georgios Lukes-Gerakopoulos,
Vojtěch Witzany
Abstract:
In this work the dynamics of a spinning particle moving in the Schwarzschild background is studied. In particular, the methods of Poincaré section and recurrence analysis are employed to discern chaos from order. It is shown that the chaotic or regular nature of the orbital motion is reflected on the gravitational waves.
In this work the dynamics of a spinning particle moving in the Schwarzschild background is studied. In particular, the methods of Poincaré section and recurrence analysis are employed to discern chaos from order. It is shown that the chaotic or regular nature of the orbital motion is reflected on the gravitational waves.
△ Less
Submitted 1 March, 2019;
originally announced March 2019.
-
Chaotic motion in the Johannsen-Psaltis spacetime
Authors:
Ondřej Zelenka,
Georgios Lukes-Gerakopoulos
Abstract:
The Johannsen-Psaltis spacetime is a perturbation of the Kerr spacetime designed to avoid pathologies like naked singularities and closed timelike curves. This spacetime depends not only on the mass and the spin of the compact object, but also on extra parameters, making the spacetime deviate from Kerr; in this work we consider only the lowest order physically meaningful extra parameter. We use nu…
▽ More
The Johannsen-Psaltis spacetime is a perturbation of the Kerr spacetime designed to avoid pathologies like naked singularities and closed timelike curves. This spacetime depends not only on the mass and the spin of the compact object, but also on extra parameters, making the spacetime deviate from Kerr; in this work we consider only the lowest order physically meaningful extra parameter. We use numerical examples to show that geodesic motion in this spacetime can exhibit chaotic behavior. We study the corresponding phase space by using Poincaré sections and rotation numbers to show chaotic behavior, and we use Lyapunov exponents to directly estimate the sensitivity to initial conditions for chaotic orbits.
△ Less
Submitted 8 November, 2017; v1 submitted 7 November, 2017;
originally announced November 2017.
