A Caltech Library Service

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

Kim, Kyungmin and Li, Tjonnie G. F. and Lo, Rico K. L. and Sachdev, Surabhi and Yuen, Robin S. H. (2020) Ranking candidate signals with machine learning in low-latency searches for gravitational waves from compact binary mergers. Physical Review D, 101 (8). Art. No. 083006. ISSN 2470-0010. doi:10.1103/physrevd.101.083006.

[img] PDF - Published Version
See Usage Policy.

[img] PDF - Submitted Version
See Usage Policy.


Use this Persistent URL to link to this item:


In the multimessenger 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, study the feasibility of adopting a supervised machine learning (ML) method for scoring rank on 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.

Item Type:Article
Related URLs:
URLURL TypeDescription Paper
Kim, Kyungmin0000-0003-1653-3795
Li, Tjonnie G. F.0000-0003-4297-7365
Sachdev, Surabhi0000-0002-0525-2317
Additional Information:© 2020 American Physical Society. Received 15 December 2019; accepted 11 March 2020; published 7 April 2020. The authors thank the LIGO Scientific Collaboration for the use of the mock data for GW150914 and corresponding background data. K. K. would like to specially thank to J. J. Oh, S. H. Oh, E. J. Son, Y.-M. Kim, W. Kim, J. Lee, S. Caudill, and K. Cannon for fruitful and constructive discussions. The work described in this paper was partially supported by grants from the Research Grants Council of the Hong Kong (Projects No. CUHK 14310816 and No. CUHK 24304317) and the Direct Grant for Research from the Research Committee of the Chinese University of Hong Kong. The work of K. K. was also supported by the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (NRF-2020R1C1C1005863).
Funding AgencyGrant Number
Research Grants Council of Hong KongCUHK 14310816
Research Grants Council of Hong KongCUHK 24304317
Chinese University of Hong KongUNSPECIFIED
National Research Foundation of KoreaNRF-2020R1C1C1005863
Issue or Number:8
Record Number:CaltechAUTHORS:20200409-092214129
Persistent URL:
Usage Policy:No commercial reproduction, distribution, display or performance rights in this work are provided.
ID Code:102427
Deposited By: Tony Diaz
Deposited On:09 Apr 2020 16:53
Last Modified:16 Nov 2021 18:12

Repository Staff Only: item control page