Ranking candidate signals with machine learning in low-latency searches for gravitational waves from compact binary mergers

Abstract

In the multi-messenger astronomy era, accurate sky localization and low latency time of gravitational-wave (GW) searches are keys in triggering successful follow-up observations on the electromagnetic counterpart of GW signals. We, in this work, focus on the latency time and study the feasibility of adopting supervised machine learning (ML) method for ranking candidate GW events. We consider two popular ML methods, random forest and neural networks. We observe that the evaluation time of both methods takes tens of milliseconds for ∼ 45,000 evaluation samples. We compare the classification efficiency between the two ML methods and a conventional low-latency search method with respect to the true positive rate at given false positive rate. The comparison shows that about 10% improved efficiency can be achieved at lower false positive rate ∼ 2 × 10 −5 with both ML methods. We also present that the search sensitivity can be enhanced by about 18% at ∼ 10 −11 Hz false alarm rate. We conclude that adopting ML methods for ranking candidate GW events is a prospective approach to yield low latency and high efficiency in searches for GW signals from compact binary mergers. PACS numbers: 95.85.Sz, 98.70.Rz, 07.05.Mh

Birth, Annexation, and Squeezing of Cities in a Prefecture: Can the Ranking of Competitive Areas of Municipalities Obey the Authentic Power Law?

“…(C) q vs. common logarithm of population density (km −2 geometry, geography, demography, linguistics, and sciences on social phenomena [35][36][37][38][39][40][41][42][43]. More recently has ranking been regarded as a tool useful for condensing large-scale data that have been accumulating in contemporary sciences such as, e.g., computational metallurgy [44] and gravitational wave astronomy [45], though the results are not yet ready for finding a rule.…”

An Attempt to Appreciate Climate Change Impacts From a Rank-Size Rule Perspective

“…To date, sustained efforts have been made to find nontrivial rules in the ranking of a variety of complex systems, not only in linguistics but in geometry, geography, demography, and sciences on social phenomena [20][21][22][23][24][25][26][36][37][38]. More recently has ranking been regarded as a tool useful for condensing largescale data that have been accumulating in contemporary sciences such as, e.g., computational metallurgy [42] and gravitational wave astronomy [43], though the results are not yet ready for finding a rule. Below, to illustrate the rank-size rule, three examples of the preceding analysis are selected in Figure 2: 1) The Metropolis of Tokyo with the entire area 2,187 km 2 consists of 62 municipalities, nine of which are located off the main land [44].…”

GWSkyNet: A Real-time Classifier for Public Gravitational-wave Candidates

“…In this work we introduce GWSkyNet, a low-latency classifier to separate real astrophysical events from terrestrial artifacts, supplementing information provided by the LIGO Scientific and Virgo Collaborations. Contrary to recent GW classifiers (Chatterjee et al 2020;Kim et al 2020), which rely on access to LIGO-Virgo data, GWSkyNet only requires publicly available data products and hence can be used by any astronomer. Built with state-of-the-art machine-learning algorithms such as non-sequential convolutional neural networks and multiple input machine-learning models, GWSkyNet achieves an accuracy of 93.5% on a testing data set.…”