曹周键,王赫,朱建阳
Cao Zhoujian,Wang He,Zhu Jianyang

• 1:北京师范大学天文系
• 2:北京师范大学物理系

摘要(Abstract)：

截至2018-01-16,LIGO已成功探测引力波事件6次.可以预期,引力波探测事件会越来越多,引力波天文学会很快进入到大数据阶段.深度学习在大数据处理方面近年来得到迅速发展.它在数据处理速度,准确度等方面都表现出极大的优势.深度学习在引力波数据处理中的应用讨论还不多.本文引入此问题,并对其进行初步研究.引力波数据最大的特点是强噪声、弱信号.现行的数据处理方法是利用匹配滤波的方式把引力波信号从强噪声中挖掘出来.同时,匹配滤波方法还可以确定引力波源的性质,定量确定其参数.匹配滤波方法的弱点是计算量巨大.这导致数据处理速度很慢.对于将来的大数据引力波天文学,这更将是一个巨大的隐患.匹配滤波方法的另一个潜在问题是,完备准确的理论波形模板是其工作的前提条件.这个潜在问题的后果是很难找到理论预期之外的引力波信号.深度学习的数据处理方法有可能在这些问题上提供出路.同时,深度学习也会遇到其自身的若干困难和问题.本文将从网络结构、训练数据制备、训练优化、对信号识别的泛化能力、对数据的特征图表示以及对特征数据遮挡的响应等方面来展开讨论.
Till 2018-01-16,LIGO has successfully detected 6 gravitational wave events.It is expected that there will be more and more gravitational wave events be detected,and the gravitational wave astronomy will quickly start its big data phase.Deep learning has developed rapidly in recent years especially on big data processing.It has shown great advantages in the speed of data processing,accuracy and so on.There is few discussion about the application of deep learning to gravitational wave data analysis.On the one hand,the current article will introduce this problem.On the other hand,we will give a preliminary exploration of this problem.The most outstanding character of gravitational wave data is weak signal hided in strong noise.The current method of gravitational wave data analysis is using the matched filtering to excavate the gravitational wave signal from the strong noise.At the same time,the properties of the gravitational wave source can be determined by the matched filtering method,and the parameters of the source can be determined quantitatively.The weakness of the matched filtering method is that the amount of computation cost is huge and the speed of data processing is slow.In the face of future big data gravitational wave astronomy,this will be a great hidden challenge.Another potential problem of the matched filtering method is that a complete and accurate theoretical waveform template is a prerequisite making sure the matched filtering technique works.The consequence of this potential problem is that it is difficult for us to find the gravitational wave signals which is out of theoretical expectation.The data processing methods of deep learning can provide a way for solving these problems.At the same time,deep learning will encounter some of its own difficulties and problems.This article will discuss the aspects of network structure,training data preparation,training optimization,generalization ability of signal recognition,representation of data character and response to feature data occlusion.

关键词(KeyWords)： 引力波天文学;匹配滤波法;深度学习方法;数值相对论;引力波形模板
gravitational wave astronomy;matched filtering method;deep learning;numerical relativity;gravitational waveform template

Abstract:

Keywords:

基金项目(Foundation): 国家自然科学基金(11622546;11690023)

作者(Author): 曹周键,王赫,朱建阳
Cao Zhoujian,Wang He,Zhu Jianyang

DOI: 10.16366/j.cnki.1000-2367.2018.02.005

参考文献(References)：

• [1]Abbott B P,Abbott R,Abbott T D,et al.Observation of Gravitational Waves from a Binary Black Hole Merger[J].Phys Rev Lett,2016,116(6):061102.
• [2]Abbott B P,Abbott R,Abbott T D,et al.GW151226:Observation of Gravitational Waves from a 22-Solar-Mass Binary Black Hole Coalescence[J].Phys Rev Lett,2016,116(24):241103.
• [3]Abbott B P,Abbott R,Abbott T D,et al.GW170608:Observation of a 19Solar-mass Binary Black Hole Coalescence[J].Ap J,2017,851:L35.
• [4]Abbott B P,Abbott R,Abbott T D,et al.GW170104:Observation of a 50-Solar-Mass Binary Black Hole Coalescence at Redshift 0.2[J].Phys Rev Lett,2017,118(22):221101.
• [5]Abbott B P,Abbott R,Abbott T D,et al.GW170814:A Three-Detector Observation of Gravitational Waves from a Binary Black Hole Coalescence[J].Phys Rev Lett,2017,119(14):141101.
• [6]Collaboration V.GW170817:Observation of Gravitational Waves from a Binary Neutron Star Inspiral[J].Phys Rev Lett,2017,119(16):161101.
• [7]Collaboration V.GW150914:First results from the search for binary black hole coalescence with Advanced LIGO[J].Phys Rev D,2016,93:122003.
• [8]Gibney E.Successful test drive for space-based gravitational-wave detector[J].Nature,2016,531(7592):20.
• [9]Gong X,Xu S,Bai S,et al.A scientific case study of an advanced LISA mission[J].Classical&Quantum Gravity,2011,28(9):192-198.
• [10]David B,Li J U,Zhao C N,et al.Gravitational wave astronomy:the current status[J].Science China Physics Mechanics&Astronomy,2015,58(12):120402.
• [11]Luo J,Chen L S,Duan H Z,et al.TianQin:a space-borne gravitational wave detector[J].Classical&Quantum Gravity,2015,33(3):035010.
• [12]蔡荣根,曹周键,韩文标.并合双星系统的引力波理论模型[J].科学通报,2016,61(14):1525-1535.
• [13]曹周键,都志辉.数值相对论与引力波天文学[J].中国科学:物理学力学天文学,2017,47(1):60-77.
• [14]CAI Ronggen,CAO Zhoujian,GUO Zongkuan,et al.The gravitational-wave physics[J].National Science Review,2017,4(5):687-706.
• [15]PRETORIUS F.Evolution of binary black-hole spacetimes.[J].Phys Rev Lett,2005,95(12):121101.
• [16]BAKER J G,CENTRELLA J,CHOI D I,et al.Gravitational-wave extraction from an inspiraling configuration of merging black holes[J].Phys Rev Lett,2006,96(11):111102.
• [17]CAMPANELLI M,LOUSTO C O,MARRONETTI P,et al.Accurate evolutions of orbiting black-hole binaries without excision[J].Phys Rev Lett,2006,96(11):111101.
• [18]BRUEGMANN B,GONZALEZ J A,HANNAM M,et al.Calibration of moving puncture simulations[J].Phys Rev D,2008,77:024027.
• [19]YAMAMOTO T,SHIBATA M,TANIGUCHI K.Simulating coalescing compact binaries by a new code sacra[J].Phys Rev D,78:064054.
• [20]CAO Z J,YO H J,YU J P.Reinvestigation of moving punctured black holes with a new code[J].Phys Rev D,2008,78:124011.
• [21]曹周键.基于爱因斯坦方程的相对论性天体模拟软件:2015SR047789[P].2015-03-18.
• [22]曹周键.从引力波探测到包含引力波的多信使天文学[Z].大学物理,2018(出版中).
• [23]Ezquiaga J M,Zumalacárregui M.Dark Energy After GW170817:Dead Ends and the Road Ahead[J].Phys Rev Lett,2017,119:251304.
• [24]Sakstein J,Jain B.Implications of the Neutron Star Merger GW170817for Cosmological Scalar-Tensor Theories[J].Phys Rev Lett,2017,119(25):251303.
• [25]Creminelli P,Vernizzi F.Dark energy after GW170817and GRB170817A[J].Phys Rev Lett,2017,119(25):251302.
• [26]Baker T,Bellini E,Ferreira P,et al.Strong Constraints on Cosmological Gravity from GW170817and GRB 170817A[J].Phys Rev Lett,2017,119(25):251301.
• [27]Vincent P,Larochelle H,Bengio Y,et al.Extracting and composing robust features with denoising autoencoders[C]//International Conference on Machine Learning.[S.l.]:ACM,2008:1096-1103.
• [28]Arbib M A.The Handbook of Brain Theory and Neural Networks[M].Cambridge:MIT Press,1998:255-258.
• [29]Silver D,Huang A,Maddison C J,et al.Mastering the game of Go with deep neural networks and tree search.[J].Nature,2016,529(7587):484.
• [30]George D,Huerta E A.Deep Neural Networks to Enable Real-time Multimessenger Astrophysics[EB/OL].[2018-01-16].https://arxiv.org/abs/1701.00008.
• [31]George D,Shen H,Huerta E A.Deep Transfer Learning:A new deep learning glitch classification method for advanced LIGO[EB/OL].[2018-01-16].https://arxiv.org/abs/1706.07446.
• [32]George D,Huerta E A.Deep Learning for Real-time Gravitational Wave Detection and Parameter Estimation:Results with Advanced LIGO Data[EB/OL].[2018-01-16].https://arxiv.org/abs/1711.03121.
• [33]George D,Shen H,Huerta E A.Glitch Classification and Clustering for LIGO with Deep Transfer Learning[EB/OL].[2018-01-16].https://arxiv.org/abs/1711.07468.
• [34]George D,Huerta E A.Deep Learning for Real-time Gravitational Wave Detection and Parameter Estimation:Results with Advanced LIGO Data[EB/OL].[2018-01-16].https://arxiv.org/abs/1711.07966.
• [35]Shen H,George D,Huerta E A,et al.Denoising Gravitational Waves using Deep Learning with Recurrent Denoising Autoencoders[EB/OL].[2018-01-16].https://arxiv.org/abs/1711.09919.
• [36]Li X,Yu W,Fan X.A Method Of Detecting Gravitational Wave Based On Time-frequency Analysis And Convolutional Neural Networks[EB/OL].[2018-01-16].https://arxiv.org/abs/1712.00356.
• [37]Cao Z,Han W B.Waveform model for an eccentric binary black hole based on the effective-one-body-numerical-relativity formalism[J].Phys Rev D,2017:96,044028.
• [38]Chen Y,Cao Z,Wang S,et al.Self-organized correlations lead to explosive synchronization[J].Physical Review E Statistical Nonlinear&Soft Matter Physics,2015,91(2):022810.
• [39]Dahl J V,Koch K C,Kleinhans E,et al.Convolutional networks and applications in vision[C]//IEEE International Symposium on Circuits and Systems.Piscataway:IEEE Press,2010:253-256.
• [40]Lecun Y,Bottou L,Bengio Y,et al.Gradient-based learning applied to document recognition[J].Proceedings of the IEEE,1998,86(11):2278-2324.
• [41]Dumoulin V,Visin F.A guide to convolution arithmetic for deep learning[EB/OL].[2018-01-16].https://arxiv.org/abs/1603.07285.
• [42]Szegedy C,Liu W,Jia Y,et al.Going deeper with convolutions[C]//IEEE Conference on Computer Vision and Pattern Recognition.Washington DC:IEEE Computer Society,2015:1-9.
• [43]Simonyan K,Zisserman A.Very Deep Convolutional Networks for Large-Scale Image Recognition[EB/OL].[2018-01-16].https://arxiv.org/abs/1409.1556.
• [44]He K,Zhang X,Ren S,et al.Deep Residual Learning for Image Recognition[EB/OL].[2018-01-16].https://arxiv.org/abs/1512.03385.
• [45]Goodfellow I,Bengio Y,Courville A.Deep Learning[M].Cambridge:MIT Press,2016.
• [46]Krizhevsky A,Sutskever I,Hinton G E.ImageNet classification with deep convolutional neural networks[C]//International Conference on Neural Information Processing Systems.[S.l.]:Curran Associates Inc,2012:1097-1105.
• [47]Glorot X,Bordes A,Bengio Y.Deep Sparse Rectifier Neural Networks[J].Jmlr W&Cp,2011,15:315-323.
• [48]Jarrett K,Kavukcuoglu K,Ranzato M,et al.What is the best multi-stage architecture for object recognition?[C]//International Conference on Computer Vision.Piscataway:IEEE Press,2010:2146-2153.
• [49]Baldi P,Sadowski P,Whiteson D.Searching for exotic particles in high-energy physics with deep learning[J].Nature Communications,2014,5(5):4308.
• [50]Lecun Y,Bengio Y,Hinton G.Deep learning[J].Nature,2015,521(7553):436-444.
• [51]Ioffe S,Szegedy C.Batch normalization:accelerating deep network training by reducing internal covariate shift[C]//International Conference on International Conference on Machine Learning.[S.l.]:JMLR,2015:448-456.
• [52]Srivastava N,Hinton G,Krizhevsky A,et al.Dropout:a simple way to prevent neural networks from overfitting[J].Journal of Machine Learning Research,2014,15(1):1929-1958.
• [53]Yu F,Koltun V.Multi-Scale Context Aggregation by Dilated Convolutions[EB/OL].[2018-01-16].https://arxiv.org/abs/1511.07122.
• [54]Yu F,Koltun V,Funkhouser T.Dilated Residual Networks[EB/OL].[2018-01-16].https://arxiv.org/abs/1705.09914.
• [55]Schmidhuber J.Deep Learning in neural networks:An overview[J].Neural Networks,2015,61:85.
• [56]Ruder S.An overview of gradient descent optimization algorithms[EB/OL].[2018-01-16].https://arxiv.org/abs/1609.04747.
• [57]Chen T,Li M,Li Y,et al.MXNet:A Flexible and Efficient Machine Learning Library for Heterogeneous Distributed Systems[EB/OL].[2018-01-16].https://arxiv.org/abs/1512.01274.
• [58]云GPU平台[EB/OL].[2018-01-16].https://www.floydhub.com.
• [59]Vidal R,Bruna J,Giryes R,et al.Mathematics of Deep Learning[EB/OL].[2018-01-16].https://arxiv.org/abs/1712.04741.
• [60]Bau D,Zhou B,Khosla A,et al.Network Dissection:Quantifying Interpretability of Deep Visual Representations[EB/OL].[2018-01-16].https://arxiv.org/abs/1704.05796.
• [61]Zeiler M D,Fergus R.Visualizing and Understanding Convolutional Networks[EB/OL].[2018-01-16].https://arxiv.org/abs/1311.2901.
• [62]Ren S,He K,Girshick R,et al.Faster R-CNN:towards real-time object detection with region proposal networks[C]//International Conference on Neural Information Processing Systems.Cambridge:MIT Press,2015:91-99.
• [63]Dai J,Li Y,He K,et al.R-FCN:Object Detection via Region-based Fully Convolutional Networks[EB/OL].[2018-01-16].https://arxiv.org/abs/1605.06409.
• [64]Liu W,Anguelov D,Erhan D,et al.SSD:Single Shot MultiBox Detector[M]//Computer Vision–ECCV 2016.Springer International Publishing,2016:21-37.
• [65]Cao Z J.Gravitational wave astronomy:chance and challenge to fundamental physics and astrophysics[J].Science China Physics Mechanics&Astronomy,2016,59(11):110431.