Download The

Total Page:16

File Type:pdf, Size:1020Kb

Download The Applying Convolutional Neural Networks to Classify Fast Radio Bursts Detected by The CHIME Telescope by Prateek Yadav A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF Master of Science in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Physics) The University of British Columbia (Vancouver) April 2020 c Prateek Yadav, 2020 The following individuals certify that they have read, and recommend to the Fac- ulty of Graduate and Postdoctoral Studies for acceptance, the thesis entitled: Applying Convolutional Neural Networks to Classify Fast Radio Bursts Detected by The CHIME Telescope submitted by Prateek Yadav in partial fulfillment of the requirements for the de- gree of Master of Science in Physics. Examining Committee: Dr. Ingrid H. Stairs, Astronomy Supervisor Dr. Gary F. Hinshaw, Astronomy Supervisory Committee Member ii Abstract The Canadian Hydrogen Intensity Mapping Experiment (CHIME) is a novel ra- dio telescope that is predicted to detect up to several dozens of Fast Radio Bursts (FRBs) per day. However, CHIME’s FRB detection software pipeline is suscep- tible to a large number of false positive triggers from terrestrial sources of Radio Frequency Interference (RFI). This thesis details the description of intensityML, a software pipeline designed to generate waterfall plots and automatically classify ra- dio bursts detected by CHIME without explicit RFI-masking and DM-refinement. The pipeline uses a convolutional neural network based classifier trained exclu- sively on the events detected by CHIME, and the classifier has an accuracy, preci- sion and recall of over 99%. It has also successfully discovered several FRBs, both in real-time and from archival data. The ideas presented in this thesis may play a key role in designing future machine-learning models for FRB classification. iii Lay Summary Fast radio bursts are bright bursts of radio waves that last about a few milliseconds in duration. These radio bursts originate from outside of our galaxy, but their exact origins remain a mystery. CHIME is a novel radio telescope which aims to detect a large number of these radio bursts in order to better understand their origins. However, radio telescopes are extremely susceptible to picking up radio signals from terrestrial sources, such as airplanes and mobile phones. This thesis presents an automated classifier, which can look at the radio bursts detected by CHIME and tell whether they had a terrestrial or an astrophysical origin. iv Preface The work presented in this thesis was primarily done by me, with contributions from various members of The CHIME/FRB collaboration. All members of collab- oration have contributed to this thesis in some way or the other through instrument and software development or through data acquisition and verification. The description of The CHIME/FRB system in Section 1.2 has been mostly summarised from previous publications by The CHIME/FRB Collaboration [4][5]. Chapter 2 presents some basic background knowledge for understanding convolu- tional neural networks, which can be readily found in most deep-learning textbooks such as Goodfellow et al., 2016 [19]. Dr. Shriharsh Tendulkar played a significant role in supervising and providing insights into the development of the plotting routine described in Section 3.2. The scripts were primarily coded by me, with some contributions from Dr. Shriharsh Tendulkar and Mr. Chitrang Patel. These scripts extensively utilise modules from Intensity Analysis Utilities, a library developed by The CHIME/FRB collaboration to analyze intensity data from radio telescopes. The idea of DM-augmentation, described in Subsection 3.2.2, also emerged from discussions within the collabo- ration. The classifier described in Section 3.1 was developed and trained indepen- dently by me. While most of the events used were labelled by the members of the collaboration, they were also double-checked by me. Mr. Charanjot Brar and Mr. Chitrang Patel assisted me with the real-time deployment of intensityML. Dr. Ingrid Stairs supervised the overall project and assisted with editing this thesis. v Table of Contents Abstract . iii Lay Summary . iv Preface . .v Table of Contents . vi List of Tables . viii List of Figures . ix Glossary . xiii Acknowledgments . xiv 1 Introduction . .1 1.1 Fast Radio Bursts . .1 1.2 The CHIME/FRB Project . .3 1.3 Related Work: Use of Machine Learning in FRB Classification . .8 1.3.1 Hybrid Deep Neural Network . .8 1.3.2 Transfer Learning on ImageNet Models . 10 1.4 Thesis Organisation . 11 2 Introduction to Convolutional Neural Networks . 12 2.1 Feed-Forward Neural-Networks . 12 vi 2.2 Convolutional Neural Networks . 15 3 Description of intensityML ........................ 19 3.1 The FRBNet Architecture . 19 3.2 Generating Waterfall Plots . 23 3.2.1 Automated Plotting Scripts . 23 3.2.2 Data Augmentation . 24 4 Results . 28 4.1 Training . 28 4.2 Results . 29 5 Discussion . 33 5.1 Discussion and Future Work . 33 5.1.1 Discussion . 33 5.1.2 Future Work . 34 5.1.3 Science Goals . 35 5.2 Conclusion . 36 Bibliography . 38 vii List of Tables Table 4.1 Accuracy, precision, recall and F1-score computed on the test set. 29 Table 4.2 L1 SNRs and DMs for events shown in Figures 4.1 and 4.2... 30 viii List of Figures Figure 1.1 (a) Plot on the left shows the waterfall plot for an FRB after correcting for the effects from electromagnetic dispersion. (b) The plot on the right shows the same waterfall plot, but with partial dedispersion to demonstrate the effects of the quadratic dispersive delay in Equation 1.1. The masked frequency chan- nels in both plots are due to interference from the LTE band. .2 Figure 1.2 CHIME radio telescope located at The Dominion Radio Astro- physical Observatory (DRAO) in Canada. Photo taken by Mark Halpern and used with permission. .4 Figure 1.3 A schematic of CHIME’s signal path. The raw data collected by the four reflectors (red arcs) is transferred to the F- and X- Engines at a rate of 13 Tb/s. The F-Engine consists of FPGA boards to digitise and channelise the data. The X-Engine utilises a GPU cluster for Fast Fourier Transform beam-forming. The CHIME/FRB backend receives the 1024 stationary intensity beams at 1 ms cadence and 16k frequency channels. Image adapted from [4].........................5 Figure 1.4 A high-level overview of the CHIME/FRB’s software pipeline showing different stages of processing. Image adapted from [4].6 Figure 1.5 A schematic diagram showing the hybrid deep neural-network developed by Connor et al. Image adapted from [9]. In the original figure, the authors seem to have incorrectly labelled the operation on the waterfall plot as 1D convolution. .9 ix Figure 1.6 Diagram showing an example of a network architecture used by Agarwal et al.. Image adapted from [2]........... 11 Figure 2.1 Schematic diagram of a fully-connected n-layer neural-network. d ( ) m An input vector x 2 R is transformed to a vector h 1 2 R in the first hidden layer. This transformation is repeated n times k as shown in Equation 2.1. The final output vector yˆ 2 R is obtained by the transformation shown in Equation 2.3, where yˆi represents the probability for class i if a softmax function (Equation 2.4) is used. Image adapted from [19]. 13 Figure 2.2 The figures show how convolution can be used to extract the edges from an input image with a single colour channel. The red pixels in the convolution kernels represent negative num- bers, while the blue pixels represent positive numbers. 16 Figure 2.3 Schematic diagram showing the operation performed by the convolution layer. See Equation 2.8. Image adapted from [19] 17 Figure 2.4 Diagram shows the typical sequence of operations in a CNN. Image adapted from [19].................... 18 Figure 3.1 Architecture of FRBNet. The model takes a single-channel 256 × 256 pixel image as input. Layer 0 of the model con- volves the input with a fixed set of thirteen 7 × 7 kernels. The resulting thirteen-channel 250×250 image is then down-sampled with 2 × 2 max-pooling and followed with non-linear ReLU activation function. The next three layers each perform a 5×5 convolution with a stride of two, followed by a ReLU activa- tion. At the end of Layer 3, the ten-channel image is down- sampled using max-pooling to give a one-dimensional vector of size ten. Finally, a fully-connected layer performs the oper- ation in Equation 2.3 to give an output vector of size 2. Image style adapted from [23]. .................... 20 x Figure 3.2 The top left plot shows the waterfall plot of an FRB. The hori- zontal streaks in this plot are RFI contamination. The remain- ing plots show the thirteen convolution kernels and their cor- responding Layer 0 transformations on the original plot. The kernel on the top right is simply the identity kernel. The re- maining kernels show the Prewitt (left) and Sobel (right) ker- nels embedded in a 7×7 grid. These help enhance the vertical pulse shape while wiping away the horizontal RFI streaks. 21 Figure 3.3 The top left plot shows the waterfall plot of an RFI event (with no astrophysical pulse present). Similar to Figure 3.2, the re- maining plots show the thirteen convolution kernels and their corresponding Layer 0 transformations on the original plot. 22 Figure 3.4 Some examples of plots generated by the automated script for events that were classified as astrophysical by the CHIME/FRB pipeline. The top five plots are pulses from FRBs and known pulsars that were correctly classified as astrophysical by the pipeline.
Recommended publications
  • Natalia Ewelina Lewandowska
    Natalia Ewelina Lewandowska - Curriculum Vitae Haverford College Office Phone: +1 304 940 6801 370 Lancester Avenue, Haverford, PA 19041 Email: [email protected] Research Gate LinkedIn Core Professional Strengths • Correlation searches and simulations of multiwavelength pulsar data • High time resolution polarimetry studies of radio pulsar data • Physics & Astronomy teaching • Undergraduate, graduate mentoring & training • Hands-on learning • Collaboration building & diversity initiatives • Public outreach management Current Position Visiting Assistant Professor of Physics and Astronomy - Haverford College Education 2010/09 - 2015/12 PhD, Physics at the Astronomy Department/University of Wu¨rzburg (A Correlation Study of Radio Giant Pulses and Very High Energy Photons from the Crab pulsar; Advisors: K. Mannheim, D. Els¨asser) 2009/01 - 2010/04 Diploma, Physics at Hamburg Observatory/University of Hamburg (Examination of Flare Stars with the highspeed photo-polarimeter OP- TIMA; Advisors: C. Liefke, J. Schmitt, G. Kanbach) 2007/10 - 2008/10 Student Research Project at Hamburg Observatory/University of Hamburg (The morphological Characteristics of Comet 17/P Holmes; Advisors: C. Liefke, J. Schmitt) 2006/04 - 2008/11 Undergraduate Student at University of Hamburg/Germany (Major: Physics, Minor: Astronomy) 2004/08 - 2006/03 Undergraduate Student at Friedrich Schiller University Jena, Extended Education Thuringia/Germany 2014/04 - 2014/07 Attendance of course “Radio Interferometry: Methods and Science” with hands-on tutorials in AIPS
    [Show full text]
  • Binary and Millisecond Pulsars Are Included in Appendix A
    Binary and Millisecond Pulsars Duncan R. Lorimer Department of Physics West Virginia University P.O. Box 6315 Morgantown WV 26506, U.S.A. email: [email protected] http://astro.wvu.edu/people/dunc Abstract We review the main properties, demographics and applications of binary and millisecond radio pulsars. Our knowledge of these exciting objects has greatly increased in recent years, mainly due to successful surveys which have brought the known pulsar population to over 1800. There are now 83 binary and millisecond pulsars associated with the disk of our Galaxy, and a further 140 pulsars in 26 of the Galactic globular clusters. Recent highlights include the discovery of the young relativistic binary system PSR J1906+0746, a rejuvination in globular cluster pulsar research including growing numbers of pulsars with masses in excess of 1.5 M⊙, a precise measurement of relativistic spin precession in the double pulsar system and a Galactic millisecond pulsar in an eccentric (e = 0.44) orbit around an unevolved companion. arXiv:0811.0762v1 [astro-ph] 5 Nov 2008 1 1 Preamble Pulsars – rapidly rotating highly magnetised neutron stars – have many applications in physics and astronomy. Striking examples include the confirmation of the existence of gravitational ra- diation [366, 367, 368], the first extra-solar planetary system [404, 293] and the first detection of gas in a globular cluster [116]. The diverse zoo of radio pulsars currently known is summarized in Figure 1. 1527 solitary pulsars SMC Globular Binary 1 4 Cluster 1 5 8 80 LMC 531 72 14 1 Planets 1 ~20 15 Supernova Recycled Remnant Figure 1: Venn diagram showing the numbers and locations of the various types of radio pulsars known as of September 2008.
    [Show full text]
  • Arxiv:1908.05356V2 [Astro-Ph.IM] 16 Aug 2019 Gravitational Waves, Extreme Astrophysics, and Fundamental Physics with Precision Pulsar Timing, J
    Astro2020 Project White Paper The NANOGrav Program for Gravitational Waves and Fundamental Physics The North American Nanohertz Observatory for Gravitational Waves 15 July 2019 Thematic areas: Multi-messenger astronomy and astrophysics; Cosmology and fundamental physics; Formation and evolution of compact objects. Contact author: Scott Ransom (NANOGrav Chair), NRAO, [email protected] Authors: A. Brazier (Cornell), S. Chatterjee (Cornell), T. Cohen (NMT), J. M. Cordes (Cornell), M. E. DeCesar (Lafayette), P. B. Demorest (NRAO), J. S. Hazboun (UW Bothell), M. T. Lam (WVU, RIT), R. S. Lynch (GBO), M. A. McLaughlin (WVU), S. M. Ransom (NRAO), X. Siemens (OSU, UWM), S. R. Taylor (Caltech/JPL, Vanderbilt), and S. J. Vigeland (UWM) for the NANOGrav Collaboration ( 50 institutions, 100+ individuals) ∼ Related NANOGrav Science White Papers arXiv:1908.05356v2 [astro-ph.IM] 16 Aug 2019 Gravitational Waves, Extreme Astrophysics, and Fundamental Physics with Precision Pulsar Timing, J. Cordes et al. · Supermassive Black-hole Demographics & Environments with Pulsar Timing Arrays, S. Taylor et al. · Fundamental Physics with Radio Millisecond Pulsars, E. Fonseca et al. · Physics Beyond the Standard Model with Pulsar Timing Arrays, X. Siemens et al. · Multi-messenger Astrophysics with Pulsar Timing Arrays, L. Kelley et al. · The Virtues of Time and Cadence for Pulsars and Fast Transients, R. Lynch et al. · Twelve Decades: Probing the ISM from kiloparsec to sub-AU scales, D. Stinebring et al. · About the cover image: Characteristic GW strain across the full spectrum. Binary black hole sources are shown in each frequency band, probed by current (PTA and ground-based) and planned (LISA) detectors. The allowed range of the stochastic background is scaled to be compatible with the NANOGrav 2018 upper limit.
    [Show full text]
  • The Advanced GBT (A-GBT)1
    Astro2020 APC White Paper Advanced Capabilities for the Green Bank Telescope Type of Activity: Ground Based Project Space Based Project Infrastructure Activity Technology Development Activity State of the Profession Consideration Other Summary: We describe projects to advance the capabilities of the GBT in survey speed, bandwidth, interference mitigation, preservation of legacy data, and community access. These upgrades will impact the study of GW and MM astronomy, fundamental physics, FRBs, cosmology, star formation, astrochemistry, distant galaxies, and searches for technosignatures. Principal Author: Ryan S. Lynch Institution: Green Bank Observatory Email: [email protected] Phone: (304) 456-2357 Co-authors: William P. Armentrout (GBO), Amber Bonsall (GBO), Natalie Butterfield (GBO), David T. Frayer (GBO), Frank Ghigo (GBO), Tapasi Ghosh (GBO), Felix J. Lockman (GBO), Ronald J. Maddalena (GBO), Anthony Minter (GBO), Larry Morgan (GBO), Karen L. O’Neil (GBO), Christopher Salter (GBO), Andrew Seymour (GBO), Joy Skipper (GBO) Endorsers: Esteban D. Araya (Western Illinois University), Hector´ G. Arce (Yale University), Filippo D’Ammando (INAF), Trisha Ashley (Caltech; Vanderbilt University), Tracy Becker (Southwest Research Institute), Robert Benjamin (University of Wisconsin-Whitewater), Sara Cazzoli (Instituto de Astrofisica de Andalucia), Tzu-Ching Chang (JPL), Suchetana Chatterjee (Presidency University), Claudia Cicone (University of Oslo), Mark J. Claussen (NRAO), Jim Cordes (Cornell University), Martin Cordiner (NASA Goddard Space Flight Center), Filippo D’Ammando (INAF-IRA Bologna), Jeremy Darling (University of Colorado), Simon Dicker (University of Pennsylvania), James Di Francesco (National Research Council of Canada), Chuanfei Dong (Princeton University), Olivier Dore´ (JPL/Caltech), Andrew Fox (STScI), Andrew Harris (University of Maryland), Jeffrey S. Hazboun (University of Washington Bothell), Ian M.
    [Show full text]
  • Natalia Ewelina Lewandowska
    Natalia Ewelina Lewandowska - Curriculum Vitae Swarthmore College Office Phone: +1 304 940 6801 500 College Ave, Swarthmore, PA 19081 Email: [email protected] Research Gate LinkedIn Core Professional Strengths • Correlation searches and simulations of multiwavelength pulsar data • High time resolution polarimetry studies of radio pulsar data • Physics & Astronomy teaching • Undergraduate, graduate mentoring & training • Hands-on learning • Collaboration building & diversity initiatives • Public outreach management Current Position Visiting Assistant Professor of Physics - Swarthmore College Adjunct Professor of Physics and Astronomy - West Virginia University Education 2010/09 - 2015/12 PhD, Physics at the Astronomy Department/University of W¨urzburg (A Correlation Study of Radio Giant Pulses and Very High Energy Photons from the Crab pulsar; Advisors: K. Mannheim, D. Els¨asser) 2009/01 - 2010/04 Diploma, Physics at Hamburg Observatory/University of Hamburg (Examination of Flare Stars with the highspeed photo-polarimeter OP- TIMA; Advisors: C. Liefke, J. Schmitt, G. Kanbach) 2007/10 - 2008/10 Student Research Project at Hamburg Observatory/University of Hamburg (The morphological Characteristics of Comet 17/P Holmes; Advisors: C. Liefke, J. Schmitt) 2006/04 - 2008/11 Undergraduate Student at University of Hamburg/Germany (Ma- jor: Physics, Minor: Astronomy) 2004/08 - 2006/03 Undergraduate Student at Friedrich Schiller University Jena, Thuringia/Germany Extended Education 2014/04 - 2014/07 Attendance of course \Radio Interferometry:
    [Show full text]
  • Dr. Robert Ferdman
    Dr. Robert Ferdman |CurriculumVitae McGill University, Department of Physics, 3600 University Street – Montreal, QC H3A 2T8, Canada T +1 (514) 398 1912 H +1 (514) 913 2135 u +1 (514) 398 8434 • • B [email protected] Õ http://www.physics.mcgill.ca/~rferdman • Research interests Gravitational wave detection with pulsar timing Tests of general relativity with pulsar observations Formation, evolution, and populations of compact binary systems Neutron-star magnetospheres and interiors Searching for new pulsars Software and instrument development Education University of British Columbia Vancouver, Canada Ph.D., Thesis supervisor: Prof. Ingrid Stairs 2003–2008 Binary pulsar systems: evolution and fundamental physics University of British Columbia Vancouver, Canada M.Sc., Thesis supervisor: Prof. Harvey Richer 2000–2002 Searching for variability in the globular cluster Messier 4 Trent University Peterborough, Canada B.Sc., Thesis supervisor: Prof. James Jury 1995–2000 Honors thesis: Simulations of a land mine detection device Employment McGill University Montreal, Canada Research associate 2012–present Supervisor: Prof. Victoria Kaspi { Continued pulsar studies, analysis, and software development for radio and X-ray data; { Leadership roles in research collaborations. University of Manchester Manchester, United Kingdom Research associate 2010–2012 Supervisor: Prof. Benjamin Stappers { Pulsar timing studies, database development; { Member of development team for data acquisition instruments; Observatoire de Paris/Radiotélescope de Nançay
    [Show full text]
  • Binary and Millisecond Pulsars Are Included in Appendix A
    Binary and Millisecond Pulsars Duncan R. Lorimer University of Manchester Jodrell Bank Observatory Macclesfield Cheshire, SK11 9DL, U.K. email: [email protected] http://www.jb.man.ac.uk/~drl Abstract We review the main properties, demographics and applications of binary and millisecond radio pulsars. Our knowledge of these exciting objects has greatly increased in recent years, mainly due to successful surveys which have brought the known pulsar population to over 1700. There are now 80 binary and millisecond pulsars associated with the disk of our Galaxy, and a further 103 pulsars in 24 of the Galactic globular clusters. Recent highlights have been the discovery of the first ever double pulsar system and a recent flurry of discoveries in globular clusters, in particular Terzan 5. arXiv:astro-ph/0511258v1 9 Nov 2005 1 1 Introduction and Overview Pulsars – rapidly rotating highly magnetised neutron stars – have resulted in many applications in physics and astronomy. Striking examples include the confirmation of the existence of gravitational radiation [314] as predicted by general relativity [312, 313], the first detection of an extra-solar planetary system [346, 244] and the discovery of the first double-pulsar binary system [44, 198]. The diverse zoo of radio pulsars currently known is summarized in Figure 1. ~ 1700 Radio Pulsars SMC Globular Binary 1 1 Cluster 1 5 6 76 LMC 181 49 4 1 Planets 1 ~20 15 Supernova ``Recycled'' Remnant Figure 1: Venn diagram showing the numbers and locations of the various types of radio pulsars known as of January 2005. The large and small Magellanic clouds are denoted by LMC and SMC.
    [Show full text]