{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,29]],"date-time":"2026-07-29T01:16:26Z","timestamp":1785287786827,"version":"3.55.0"},"reference-count":47,"publisher":"IOP Publishing","issue":"3","license":[{"start":{"date-parts":[[2021,6,16]],"date-time":"2021-06-16T00:00:00Z","timestamp":1623801600000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2021,6,16]],"date-time":"2021-06-16T00:00:00Z","timestamp":1623801600000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/iopscience.iop.org\/info\/page\/text-and-data-mining"}],"content-domain":{"domain":["iopscience.iop.org"],"crossmark-restriction":false},"short-container-title":["Mach. Learn.: Sci. Technol."],"published-print":{"date-parts":[[2021,9,1]]},"abstract":"<jats:title>Abstract<\/jats:title>\n               <jats:p>The LIGO and Virgo gravitational-wave observatories have detected many exciting events over the past 5 years. To infer the system parameters, iterative sampling algorithms such as MCMC are typically used with Bayes\u2019 theorem to obtain posterior samples\u2014by repeatedly generating waveforms and comparing to measured strain data. However, as the rate of detections grows with detector sensitivity, this poses a growing computational challenge. To confront this challenge, as well as that of fast multimessenger alerts, in this study we apply deep learning to learn non-iterative surrogate models for the Bayesian posterior. We train a neural-network conditional density estimator to model posterior probability distributions over the full 15-dimensional space of binary black hole system parameters, given detector strain data from multiple detectors. We use the method of normalizing flows\u2014specifically, a <jats:italic>neural spline<\/jats:italic> flow\u2014which allows for rapid sampling and density estimation. Training the network is likelihood-free, requiring samples from the data generative process, but no likelihood evaluations. Through training, the network learns a <jats:italic>global<\/jats:italic> set of posteriors: it can generate thousands of independent posterior samples per second for any strain data consistent with the training distribution. We demonstrate our method by performing inference on GW150914, and obtain results in close agreement with standard techniques.<\/jats:p>","DOI":"10.1088\/2632-2153\/abfaed","type":"journal-article","created":{"date-parts":[[2021,4,22]],"date-time":"2021-04-22T23:01:44Z","timestamp":1619132504000},"page":"03LT01","update-policy":"https:\/\/doi.org\/10.1088\/crossmark-policy","source":"Crossref","is-referenced-by-count":90,"title":["Complete parameter inference for GW150914 using deep learning"],"prefix":"10.1088","volume":"2","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-6987-6313","authenticated-orcid":false,"given":"Stephen R","family":"Green","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-1671-3668","authenticated-orcid":false,"given":"Jonathan","family":"Gair","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"266","published-online":{"date-parts":[[2021,6,16]]},"reference":[{"key":"mlstabfaedbib1","doi-asserted-by":"publisher","DOI":"10.1103\/PhysRevLett.116.061102","article-title":"Observation of gravitational waves from a binary black hole merger","volume":"116","author":"Abbott","year":"2016","journal-title":"Phys. Rev. Lett."},{"key":"mlstabfaedbib2","doi-asserted-by":"publisher","DOI":"10.1103\/PhysRevX.9.031040","article-title":"GWTC-1: A Gravitational-Wave Transient Catalog of Compact Binary Mergers Observed by LIGO and Virgo during the First and Second Observing Runs","volume":"9","author":"Abbott","year":"2019","journal-title":"Phys. Rev. X"},{"key":"mlstabfaedbib3","doi-asserted-by":"publisher","DOI":"10.1103\/PhysRevD.102.043015","article-title":"GW190412: observation of a binary-black-hole coalescence with asymmetric masses","volume":"102","author":"Abbott","year":"2020"},{"key":"mlstabfaedbib4","doi-asserted-by":"publisher","first-page":"L3","DOI":"10.3847\/2041-8213\/ab75f5","article-title":"GW190425: observation of a compact binary coalescence with total mass \u20093.4M\u2299","volume":"892","author":"Abbott","year":"2020","journal-title":"Astrophys. J. Lett."},{"key":"mlstabfaedbib5","doi-asserted-by":"publisher","first-page":"L44","DOI":"10.3847\/2041-8213\/ab960f","article-title":"GW190814: gravitational waves from the coalescence of a 23 solar mass black hole with a 2.6 solar mass compact object","volume":"896","author":"Abbott","year":"2020","journal-title":"Astrophys. J."},{"key":"mlstabfaedbib6","article-title":"GWTC-2: compact binary coalescences observed by LIGO and Virgo during the first half of the third observing run","author":"Abbott","year":"2020"},{"key":"mlstabfaedbib7","article-title":"Gravitational-wave candidate event database"},{"key":"mlstabfaedbib8","doi-asserted-by":"publisher","DOI":"10.1103\/PhysRevD.91.042003","article-title":"Parameter estimation for compact binaries with ground-based gravitational-wave observations using the LALInference software library","volume":"D91","author":"Veitch","year":"2015","journal-title":"Phys. Rev."},{"key":"mlstabfaedbib9","doi-asserted-by":"publisher","first-page":"3295","DOI":"10.1093\/mnras\/staa2850","article-title":"Bayesian inference for compact binary coalescences with BILBY: validation and application to the first LIGO\u2013Virgo gravitational-wave transient catalogue","volume":"499","author":"Romero-Shaw","year":"2020","journal-title":"Mon. Notices Royal Astron. Soc"},{"key":"mlstabfaedbib10","doi-asserted-by":"publisher","DOI":"10.1103\/PhysRevX.9.011001","article-title":"Properties of the binary neutron star merger GW170817","volume":"X9","author":"Abbott","year":"2019","journal-title":"Phys. Rev."},{"key":"mlstabfaedbib11","doi-asserted-by":"publisher","first-page":"3","DOI":"10.1007\/s41114-020-00026-9","article-title":"Prospects for observing and localizing gravitational-wave transients with advanced LIGO, advanced Virgo and KAGRA","volume":"21","author":"Abbott","year":"2018","journal-title":"Living Rev. Rel."},{"key":"mlstabfaedbib12","doi-asserted-by":"publisher","first-page":"64","DOI":"10.1016\/j.physletb.2017.12.053","article-title":"Deep learning for real-time gravitational wave detection and parameter estimation: results with advanced LIGO data","volume":"778","author":"George","year":"2018","journal-title":"Phys. Lett. B"},{"key":"mlstabfaedbib13","article-title":"Deep learning for multi-messenger astrophysics: a gateway for discovery in the big data era","author":"Allen","year":"2019"},{"key":"mlstabfaedbib14","doi-asserted-by":"publisher","DOI":"10.1088\/2632-2153\/abb93a","article-title":"Enhancing gravitational-wave science with machine learning","volume":"2","author":"Cuoco","year":"2020","journal-title":"Mach. Learn.: Sci. Technol."},{"key":"mlstabfaedbib15","doi-asserted-by":"publisher","DOI":"10.1103\/PhysRevLett.124.041102","article-title":"Learning Bayesian posteriors with neural networks for gravitational-wave inference","volume":"124","author":"Chua","year":"2020","journal-title":"Phys. Rev. Lett."},{"key":"mlstabfaedbib16","article-title":"Bayesian parameter estimation using conditional variational autoencoders for gravitational-wave astronomy","author":"Gabbard","year":"2019"},{"key":"mlstabfaedbib17","doi-asserted-by":"publisher","DOI":"10.1103\/PhysRevD.102.104057","article-title":"Gravitational-wave parameter estimation with autoregressive neural network flows","volume":"102","author":"Green","year":"2020","journal-title":"Phys. Rev. D"},{"key":"mlstabfaedbib18","article-title":"Auto-encoding variational Bayes","author":"Kingma","year":"2013"},{"key":"mlstabfaedbib19","first-page":"1278","article-title":"Stochastic backpropagation and approximate inference in deep generative models","author":"Rezende","year":"2014"},{"key":"mlstabfaedbib20","first-page":"1530","article-title":"Variational inference with normalizing flows","author":"Rezende","year":"2015"},{"key":"mlstabfaedbib21","first-page":"4743","article-title":"Improved variational inference with inverse autoregressive flow","author":"Kingma","year":"2016"},{"key":"mlstabfaedbib22","first-page":"2338","article-title":"Masked autoregressive flow for density estimation","author":"Papamakarios","year":"2017"},{"key":"mlstabfaedbib23","article-title":"Variational Lossy autoencoder","author":"Chen","year":"2016"},{"key":"mlstabfaedbib24","article-title":"Advanced LIGO anticipated sensitivity curves","year":"2009"},{"key":"mlstabfaedbib25","article-title":"Neural spline flows","author":"Durkan","year":"2019"},{"key":"mlstabfaedbib26","article-title":"Identity mappings in deep residual networks","author":"Kaiming","year":"2016"},{"key":"mlstabfaedbib27","first-page":"448","article-title":"Batch normalization: accelerating deep network training by reducing internal covariate shift","author":"Ioffe","year":"2015"},{"key":"mlstabfaedbib28","article-title":"Fast and accurate deep network learning by exponential linear units (elus)","author":"Clevert","year":"2015"},{"key":"mlstabfaedbib29","article-title":"Neural spline flows","author":"Durkan","year":"2019"},{"key":"mlstabfaedbib30","doi-asserted-by":"publisher","DOI":"10.1103\/PhysRevD.90.024018","article-title":"A more effective coordinate system for parameter estimation of precessing compact binaries from gravitational waves","volume":"90","author":"Farr","year":"2014","journal-title":"Phys. Rev. D"},{"key":"mlstabfaedbib31","article-title":"Nice: non-linear independent components estimation","author":"Dinh","year":"2014"},{"key":"mlstabfaedbib32","article-title":"Deep residual learning for image recognition","author":"Kaiming","year":"2015"},{"key":"mlstabfaedbib33","article-title":"Adam: a method for stochastic optimization","author":"Kingma","year":"2014"},{"key":"mlstabfaedbib34","doi-asserted-by":"publisher","DOI":"10.1103\/PhysRevLett.113.151101","article-title":"Simple model of complete precessing black-hole-binary gravitational waveforms","volume":"113","author":"Hannam","year":"2014","journal-title":"Phys. Rev. Lett."},{"key":"mlstabfaedbib35","doi-asserted-by":"publisher","DOI":"10.1103\/PhysRevD.93.044007","article-title":"Frequency-domain gravitational waves from nonprecessing black-hole binaries. II. A phenomenological model for the advanced detector era","volume":"D93","author":"Khan","year":"2016","journal-title":"Phys. Rev."},{"key":"mlstabfaedbib36","article-title":"PhenomPv2\u2014technical notes for the LAL implementation","author":"Boh\u00e9","year":"2016"},{"key":"mlstabfaedbib37","doi-asserted-by":"publisher","DOI":"10.1103\/PhysRevD.87.124005","article-title":"Gravitational wave parameter estimation with compressed likelihood evaluations","volume":"87","author":"Canizares","year":"2013","journal-title":"Phys. Rev. D"},{"key":"mlstabfaedbib38","doi-asserted-by":"publisher","DOI":"10.1103\/PhysRevD.94.044031","article-title":"Fast and accurate inference on gravitational waves from precessing compact binaries","volume":"94","author":"Rory Smith","year":"2016","journal-title":"Phys. Rev. D"},{"key":"mlstabfaedbib39","article-title":"Sgdr: stochastic gradient descent with warm restarts","author":"Loshchilov","year":"2016"},{"key":"mlstabfaedbib40","doi-asserted-by":"publisher","first-page":"27","DOI":"10.3847\/1538-4365\/ab06fc","article-title":"BILBY: a user-friendly Bayesian inference library for gravitational-wave astronomy","volume":"241","author":"Ashton","year":"2019","journal-title":"Astrophys. J. Suppl."},{"key":"mlstabfaedbib41","doi-asserted-by":"publisher","first-page":"3132","DOI":"10.1093\/mnras\/staa278","article-title":"Dynesty: a dynamic nested sampling package for estimating Bayesian posteriors and evidences","volume":"493","author":"Speagle","year":"2020","journal-title":"Mon. Not. R. Astron. Soc."},{"key":"mlstabfaedbib42","doi-asserted-by":"publisher","DOI":"10.1103\/PhysRevD.102.044055","article-title":"Multipolar effective-one-body waveforms for precessing binary black holes: construction and validation","volume":"102","author":"Ossokine","year":"2020","journal-title":"Phys. Rev. D"},{"key":"mlstabfaedbib43","doi-asserted-by":"publisher","DOI":"10.1016\/j.physletb.2021.136161","article-title":"Detection and parameter estimation of gravitational waves from binary neutron-star mergers in real LIGO data using deep learning","volume":"815","author":"Krastev","year":"2021","journal-title":"Phys. Lett. B"},{"key":"mlstabfaedbib44","first-page":"pp 8024","author":"Paszke","year":"2019"},{"key":"mlstabfaedbib45","doi-asserted-by":"publisher","first-page":"90","DOI":"10.1109\/MCSE.2007.55","article-title":"Matplotlib: a 2d graphics environment","volume":"9","author":"Hunter","year":"2007","journal-title":"Comput. Sci. Eng."},{"key":"mlstabfaedbib46","doi-asserted-by":"publisher","DOI":"10.21105\/joss.00045","article-title":"ChainConsumer","volume":"1","author":"Hinton","year":"2016","journal-title":"J. Open Source Softw."},{"key":"mlstabfaedbib47","article-title":"ligo.skymap","author":"Singer","year":"2020"}],"container-title":["Machine Learning: Science and Technology"],"original-title":[],"link":[{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/abfaed","content-type":"text\/html","content-version":"am","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/abfaed\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/abfaed","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/abfaed\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/abfaed\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"syndication"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/abfaed\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/abfaed\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"similarity-checking"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/abfaed\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2021,12,13]],"date-time":"2021-12-13T13:53:13Z","timestamp":1639403593000},"score":1,"resource":{"primary":{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/abfaed"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2021,6,16]]},"references-count":47,"journal-issue":{"issue":"3","published-online":{"date-parts":[[2021,6,16]]},"published-print":{"date-parts":[[2021,9,1]]}},"URL":"https:\/\/doi.org\/10.1088\/2632-2153\/abfaed","relation":{},"ISSN":["2632-2153"],"issn-type":[{"value":"2632-2153","type":"electronic"}],"subject":[],"published":{"date-parts":[[2021,6,16]]},"assertion":[{"value":"Complete parameter inference for GW150914 using deep learning","name":"article_title","label":"Article Title"},{"value":"Machine Learning: Science and Technology","name":"journal_title","label":"Journal Title"},{"value":"paper","name":"article_type","label":"Article Type"},{"value":"\u00a9 2021 The Author(s). Published by IOP Publishing Ltd","name":"copyright_information","label":"Copyright Information"},{"value":"2021-02-26","name":"date_received","label":"Date Received","group":{"name":"publication_dates","label":"Publication dates"}},{"value":"2021-04-22","name":"date_accepted","label":"Date Accepted","group":{"name":"publication_dates","label":"Publication dates"}},{"value":"2021-06-16","name":"date_epub","label":"Online publication date","group":{"name":"publication_dates","label":"Publication dates"}}]}}