Mathematics of Data Science

Total Page:16

File Type:pdf, Size:1020Kb

Mathematics of Data Science SIAM JOURNAL ON Mathematics of Data Science Volume 2 • 2020 Editor-in-Chief Tamara G. Kolda, Sandia National Laboratories Section Editors Mark Girolami, University of Cambridge, UK Alfred Hero, University of Michigan, USA Robert D. Nowak, University of Wisconsin, Madison, USA Joel A. Tropp, California Institute of Technology, USA Associate Editors Maria-Florina Balcan, Carnegie Mellon University, USA Vianney Perchet, ENSAE, CRITEO, France Rina Foygel Barber, University of Chicago, USA Jonas Peters, University of Copenhagen, Denmark Mikhail Belkin, University of California, San Diego, USA Natesh Pillai, Harvard University, USA Robert Calderbank, Duke University, USA Ali Pinar, Sandia National Laboratories, USA Coralia Cartis, University of Oxford, UK Mason Porter, University of Califrornia, Los Angeles, USA Venkat Chandrasekaran, California Institute of Technology, Maxim Raginsky, University of Illinois, USA Urbana-Champaign, USA Patrick L. Combettes, North Carolina State University, USA Bala Rajaratnam, University of California, Davis, USA Alexandre d’Aspremont, CRNS, Ecole Normale Superieure, Philippe Rigollet, MIT, USA France Justin Romberg, Georgia Tech, USA Ioana Dumitriu, University of California, San Diego, USA C. Seshadhri, University of California, Santa Cruz, USA Maryam Fazel, University of Washington, USA Amit Singer, Princeton University, USA David F. Gleich, Purdue University, USA Marc Teboulle, Tel Aviv University, Israel Wouter Koolen, CWI, the Netherlands Caroline Uhler, MIT, USA Gitta Kutyniok, University of Munich, Germany Ramon van Handel, Princeton University, USA Monique Laurent, CWI, the Netherlands Weichung Wang, National Taiwan University, Taiwan Yi Ma, University of California, Berkeley, USA Rachel Ward, University of Texas, Austin, USA Michael Mahoney, University of California, Berkeley, USA Rebecca Willett, University of Chicago, USA Boaz Nadler, Weizmann Institute of Science, Israel Lenka Zdeborová, CNRS, CEA/SACLAY, France Long Nguyen, University of Michigan, USA Ivan Oseledets, Skolkovo Institute of Science and Technology, Russia Articles copyright © 2020 by the Society for Industrial and Applied Mathematics, Philadelphia, PA, 19104 USA.
Recommended publications
  • Eigenvalue Distributions of Beta-Wishart Matrices
    Eigenvalue distributions of beta-Wishart matrices The MIT Faculty has made this article openly available. Please share how this access benefits you. Your story matters. Citation Edelman, Alan, and Plamen Koev. “Eigenvalue Distributions of Beta- Wishart Matrices.” Random Matrices: Theory and Applications 03, no. 02 (April 2014): 1450009. As Published http://dx.doi.org/10.1142/S2010326314500099 Publisher World Scientific Pub Co Pte Lt Version Author's final manuscript Citable link http://hdl.handle.net/1721.1/116006 Terms of Use Creative Commons Attribution-Noncommercial-Share Alike Detailed Terms http://creativecommons.org/licenses/by-nc-sa/4.0/ SIAM J. MATRIX ANAL. APPL. c 2013 Society for Industrial and Applied Mathematics Vol. XX, No. X, pp. XX{XX EIGENVALUE DISTRIBUTIONS OF BETA-WISHART MATRICES∗ ALAN EDELMANy AND PLAMEN KOEVz Abstract. We derive explicit expressions for the distributions of the extreme eigenvalues of the Beta-Wishart random matrices in terms of the hypergeometric function of a matrix argument. These results generalize the classical results for the real (β = 1), complex (β = 2), and quaternion (β = 4) Wishart matrices to any β > 0. Key words. random matrix, Wishart distribution, eigenavalue, hypergeometric function of a matrix argument AMS subject classifications. 15A52, 60E05, 62H10, 65F15 DOI. XX.XXXX/SXXXXXXXXXXXXXXXX 1. Introduction. Recently, the classical real (β = 1), complex (β = 2), and quaternion (β = 4) Wishart random matrix ensembles were generalized to any β > 0 by what is now called the Beta{Wishart ensemble [2, 9]. In this paper we derive the explicit distributions for the extreme eigenvalues and the trace of this ensemble as series of Jack functions and, in particular, in terms of the hypergeometric function of matrix argument.
    [Show full text]
  • Matrix Algorithms: Fast, Stable, Communication-Optimizing—Random?!
    Matrix algorithms: fast, stable, communication-optimizing—random?! Ioana Dumitriu Department of Mathematics University of Washington (Seattle) Joint work with Grey Ballard, James Demmel, Olga Holtz, Robert Kleinberg IPAM: Convex Optimization and Algebraic Geometry September 28, 2010 Ioana Dumitriu (UW) Matrix algorithms September 28, 2010 1 / 49 1 Motivation and Background Main Concerns in Numerical Linear Algebra Size issues 2 Reducing the flops Stability of FMM Stability of FLA Why Randomize? 3 Reducing communication Why is communication bad? Randomized Spectral Divide and Conquer Communication Costs 4 Conclusions Ioana Dumitriu (UW) Matrix algorithms September 28, 2010 2 / 49 Motivation and Background Main Concerns in Numerical Linear Algebra The big four issues in matrix algorithms. Accuracy of computation (how “far” the computed values are from the actual ones, in the presence of rounding error). Stability (how much the computed result will change if we perturb the problem a little bit). Speed/complexity (how many operations they require, e.g., multiplications). Parallelism (how to distribute the computation to multiple processors in order to optimize speed – if possible). Communication complexity (how to minimize the amount of back-and-forth between levels of memory or processors.) Ioana Dumitriu (UW) Matrix algorithms September 28, 2010 3 / 49 Motivation and Background Main Concerns in Numerical Linear Algebra The big four issues in matrix algorithms. Accuracy of computation (how “far” the computed values are from the actual ones, in the presence of rounding error). Stability (how much the computed result will change if we perturb the problem a little bit). Speed/flops complexity (how many operations they require, e.g., multiplications).
    [Show full text]
  • The Random Matrix Technique of Ghosts and Shadows
    The Random Matrix Technique of Ghosts and Shadows Alan Edelman November 22, 2009 Abstract We propose to abandon the notion that a random matrix has to be sampled for it to exist. Much of today's applied nite random matrix theory concerns real or complex random matrices (β = 1; 2). The threefold way so named by Dyson in 1962 [2] adds quaternions (β = 4). While it is true there are only three real division algebras (β=dimension over the reals), this mathematical fact while critical in some ways, in other ways is irrelevant and perhaps has been over interpreted over the decades. We introduce the notion of a ghost random matrix quantity that exists for every beta, and a shadow quantity which may be real or complex which allows for computation. Any number of computations have successfully given reasonable answers to date though diculties remain in some cases. Though it may seem absurd to have a three and a quarter dimensional or pi dimensional algebra, that is exactly what we propose and what we compute with. In the end β becomes a noisiness parameter rather than a dimension. 1 Introduction This conference article contains an idea which has become a technique. Perhaps it might be labeled a conjecture, but I think idea is the better label right now. The idea was discussed informally to a number of researchers and students at MIT for a number of years now, probably dating back to 2003 or so. It was also presented at a number of conferences [3] . As slides do not quite capture a talk, this seemed a good place to write down the ideas.
    [Show full text]
  • Digital Communication Systems 2.2 Optimal Source Coding
    Digital Communication Systems EES 452 Asst. Prof. Dr. Prapun Suksompong [email protected] 2. Source Coding 2.2 Optimal Source Coding: Huffman Coding: Origin, Recipe, MATLAB Implementation 1 Examples of Prefix Codes Nonsingular Fixed-Length Code Shannon–Fano code Huffman Code 2 Prof. Robert Fano (1917-2016) Shannon Award (1976 ) Shannon–Fano Code Proposed in Shannon’s “A Mathematical Theory of Communication” in 1948 The method was attributed to Fano, who later published it as a technical report. Fano, R.M. (1949). “The transmission of information”. Technical Report No. 65. Cambridge (Mass.), USA: Research Laboratory of Electronics at MIT. Should not be confused with Shannon coding, the coding method used to prove Shannon's noiseless coding theorem, or with Shannon–Fano–Elias coding (also known as Elias coding), the precursor to arithmetic coding. 3 Claude E. Shannon Award Claude E. Shannon (1972) Elwyn R. Berlekamp (1993) Sergio Verdu (2007) David S. Slepian (1974) Aaron D. Wyner (1994) Robert M. Gray (2008) Robert M. Fano (1976) G. David Forney, Jr. (1995) Jorma Rissanen (2009) Peter Elias (1977) Imre Csiszár (1996) Te Sun Han (2010) Mark S. Pinsker (1978) Jacob Ziv (1997) Shlomo Shamai (Shitz) (2011) Jacob Wolfowitz (1979) Neil J. A. Sloane (1998) Abbas El Gamal (2012) W. Wesley Peterson (1981) Tadao Kasami (1999) Katalin Marton (2013) Irving S. Reed (1982) Thomas Kailath (2000) János Körner (2014) Robert G. Gallager (1983) Jack KeilWolf (2001) Arthur Robert Calderbank (2015) Solomon W. Golomb (1985) Toby Berger (2002) Alexander S. Holevo (2016) William L. Root (1986) Lloyd R. Welch (2003) David Tse (2017) James L.
    [Show full text]
  • Principles of Communications ECS 332
    Principles of Communications ECS 332 Asst. Prof. Dr. Prapun Suksompong (ผศ.ดร.ประพันธ ์ สขสมปองุ ) [email protected] 1. Intro to Communication Systems Office Hours: Check Google Calendar on the course website. Dr.Prapun’s Office: 6th floor of Sirindhralai building, 1 BKD 2 Remark 1 If the downloaded file crashed your device/browser, try another one posted on the course website: 3 Remark 2 There is also three more sections from the Appendices of the lecture notes: 4 Shannon's insight 5 “The fundamental problem of communication is that of reproducing at one point either exactly or approximately a message selected at another point.” Shannon, Claude. A Mathematical Theory Of Communication. (1948) 6 Shannon: Father of the Info. Age Documentary Co-produced by the Jacobs School, UCSD- TV, and the California Institute for Telecommunic ations and Information Technology 7 [http://www.uctv.tv/shows/Claude-Shannon-Father-of-the-Information-Age-6090] [http://www.youtube.com/watch?v=z2Whj_nL-x8] C. E. Shannon (1916-2001) Hello. I'm Claude Shannon a mathematician here at the Bell Telephone laboratories He didn't create the compact disc, the fax machine, digital wireless telephones Or mp3 files, but in 1948 Claude Shannon paved the way for all of them with the Basic theory underlying digital communications and storage he called it 8 information theory. C. E. Shannon (1916-2001) 9 https://www.youtube.com/watch?v=47ag2sXRDeU C. E. Shannon (1916-2001) One of the most influential minds of the 20th century yet when he died on February 24, 2001, Shannon was virtually unknown to the public at large 10 C.
    [Show full text]
  • MIMO Wireless Communications Ezio Biglieri, Robert Calderbank, Anthony Constantinides, Andrea Goldsmith, Arogyaswami Paulraj and H
    Cambridge University Press 978-0-521-13709-6 - MIMO Wireless Communications Ezio Biglieri, Robert Calderbank, Anthony Constantinides, Andrea Goldsmith, Arogyaswami paulraj and H. Vincent Poor Frontmatter More information MIMO Wireless Communications Multiple-input multiple-output (MIMO) technology constitutes a breakthrough in the design of wireless communication systems, and is already at the core of several wireless standards. Exploiting multi-path scattering, MIMO techniques deliver significant performance enhancements in terms of data transmission rate and interference reduction. This book is a detailed introduction to the analysis and design of MIMO wireless systems. Beginning with an overview of MIMO technology, the authors then examine the fundamental capacity limits of MIMO systems. Transmitter design, including precoding and space–time coding, is then treated in depth, and the book closes with two chapters devoted to receiver design. Written by a team of leading experts, the book blends theoretical analysis with physical insights, and highlights a range of key design challenges. It can be used as a textbook for advanced courses on wireless communications, and will also appeal to researchers and practitioners working on MIMO wireless systems. Ezio Biglieri is a professor in the Department of Technology at the Universitat Pompeu Fabra, Barcelona. Robert Calderbank is a professor in the Departments of Electrical Engineering and Mathematics at Princeton University, New Jersey. Anthony Constantinides is a professor in the Department of Electrical and Electronic Engineering at Imperial College of Science, Technology and Medicine, London. Andrea Goldsmith is a professor in the Department of Electrical Engineering at Stanford University, California. Arogyaswami Paulraj is a professor in the Department of Electrical Engineering at Stanford University, California.
    [Show full text]
  • 2017-2018 Annual Report
    Institute for Computational and Experimental Research in Mathematics Annual Report May 1, 2017 – April 30, 2018 Brendan Hassett, Director Mathew Borton, IT Director Jeff Brock, Associate Director Ruth Crane, Assistant Director Sigal Gottlieb, Deputy Director Elisenda Grigsby, Deputy Director Jeffrey Hoffstein, Consulting Associate Director Caroline Klivans, Associate Director Jill Pipher, Consulting Associate Director Bjorn Sandstede, Associate Director Homer Walker, Consulting Associate Director Ulrica Wilson, Associate Director for Diversity and Outreach Table of Contents Mission ....................................................................................................................................... 6 Core Programs and Events ......................................................................................................... 6 Participant Summaries by Program Type ................................................................................... 9 ICERM Funded Participants ................................................................................................................. 9 All Participants (ICERM funded and Non-ICERM funded) .................................................................. 10 ICERM Funded Speakers ................................................................................................................... 11 All Speakers (ICERM funded and Non-ICERM funded) ...................................................................... 12 ICERM Funded Postdocs ..................................................................................................................
    [Show full text]
  • ANNOUNCEMENT of WINNERS of the Seventy-Ninth Competition Held on December 1, 2018
    ANNOUNCEMENT OF WINNERS of the seventy-ninth competition held on December 1, 2018 WINNING TEAMS Rank School Team Members (in alphabetical order) 1 HARVARD UNIVERSITY Dongryul Kim, Shyam Narayanan, David Stoner 2 MASSACHUSETTS INSTITUTE OF TECHNOLOGY Junyao Peng, Ashwin Sah, Yunkun Zhou 3 UNIVERSITY OF CALIFORNIA, LOS ANGELES Ciprian Mircea Bonciocat, Xiaoyu Huang, Konstantin Miagkov 4 COLUMBIA UNIVERSITY Quang Dao, Myeonhu Kim, Matthew Lerner-Brecher 5 STANFORD UNIVERSITY David Kewei Lin, Hanzhi Zheng, Yifan Zhu The institution with the first-place team receives an award of $25,000, and each member of the team receives $1,000. The awards for second place are $20,000 and $800; for third place, $15,000 and $600; for fourth place, $10,000 and $400; and for fifth place, $5,000 and $200. In each of the following categories, the listing is in alphabetical order. THE PUTNAM FELLOWS — THE FIVE HIGHEST RANKING INDIVIDUALS Each receives an award of $2,500. DONGRYUL KIM Harvard University SHYAM NARAYANAN Harvard University DAVID STONER Harvard University YUAN YAO Massachusetts Institute of Technology SHENGTONG ZHANG Massachusetts Institute of Technology THE NEXT TEN HIGHEST RANKING INDIVIDUALS Each receives an award of $1,000. MURILO CORATO ZANARELLA Princeton University JIYANG GAO Massachusetts Institute of Technology ANDREW GU Massachusetts Institute of Technology JAMES LIN Massachusetts Institute of Technology MICHAEL MA Massachusetts Institute of Technology ASHWIN SAH Massachusetts Institute of Technology KEVIN SUN Massachusetts Institute of Technology DANIELLE WANG Massachusetts Institute of Technology HUNG-HSUN YU Massachusetts Institute of Technology YUNKUN ZHOU Massachusetts Institute of Technology THE WILLIAM LOWELL PUTNAM MATHEMATICAL COMPETITION ELIZABETH LOWELL PUTNAM PRIZE The winner receives an award of $1,000.
    [Show full text]
  • Scientific Workplace· • Mathematical Word Processing • LATEX Typesetting Scientific Word· • Computer Algebra
    Scientific WorkPlace· • Mathematical Word Processing • LATEX Typesetting Scientific Word· • Computer Algebra (-l +lr,:znt:,-1 + 2r) ,..,_' '"""""Ke~r~UrN- r o~ r PooiliorK 1.931'J1 Po6'lf ·1.:1l26!.1 Pod:iDnZ 3.881()2 UfW'IICI(JI)( -2.801~ ""'"""U!NecteoZ l!l!iS'11 v~ 0.7815399 Animated plots ln spherical coordln1tes > To make an anlm.ted plot In spherical coordinates 1. Type an expression In thr.. variables . 2 WMh the Insertion poilt In the expression, choose Plot 3D The next exampfe shows a sphere that grows ftom radius 1 to .. Plot 3D Animated + Spherical The Gold Standard for Mathematical Publishing Scientific WorkPlace and Scientific Word Version 5.5 make writing, sharing, and doing mathematics easier. You compose and edit your documents directly on the screen, without having to think in a programming language. A click of a button allows you to typeset your documents in LAT£X. You choose to print with or without LATEX typesetting, or publish on the web. Scientific WorkPlace and Scientific Word enable both professionals and support staff to produce stunning books and articles. Also, the integrated computer algebra system in Scientific WorkPlace enables you to solve and plot equations, animate 20 and 30 plots, rotate, move, and fly through 3D plots, create 3D implicit plots, and more. MuPAD' Pro MuPAD Pro is an integrated and open mathematical problem­ solving environment for symbolic and numeric computing. Visit our website for details. cK.ichan SOFTWARE , I NC. Visit our website for free trial versions of all our products. www.mackichan.com/notices • Email: info@mac kichan.com • Toll free: 877-724-9673 It@\ A I M S \W ELEGRONIC EDITORIAL BOARD http://www.math.psu.edu/era/ Managing Editors: This electronic-only journal publishes research announcements (up to about 10 Keith Burns journal pages) of significant advances in all branches of mathematics.
    [Show full text]
  • 2009 Mathematics Newsletter
    AUTUMN 2009 NEWSLETTER OF THE DEPARTMENT OF MATHEMATICS AT THE UNIVERSITY OF WASHINGTON Mathematics NEWS 1 DEPARTMENT OF MATHEMATICS NEWS MESSAGE FROM THE CHAIR As you might imagine, 2009 brought Department’s work continued unabated. We have a string of unusual challenges to our department. good news to share with you in this newsletter, such as the There were uncertainties well into the UW Sophomore Medal awarded to Chad Klumb, the new- spring regarding our 2009-10 aca- found success of our students in the Putnam Competition, demic year budget. Graduate student the election of Gunther Uhlmann to the American Academy recruitment, which takes place across of Arts and Sciences, and the NSF CAREER award to Ioana the nation during the winter, was Dumitriu. Our undergraduate degree programs set new highs conducted without knowing the level at the end of the 2008-09 academic year with 368 Math of our TA funding. Similar scenarios majors, and with a total of 537 majors in Math and ACMS. played out nationally as few educational institutions were Thirteen graduate students completed the PhD, continuing spared by the financial crisis. In a survey conducted by the our recent trend of awarding significantly more PhDs than American Mathematical Society, mathematics departments the Department’s historical annual average of 6.5. reported about 900 expected faculty openings nationally, During the spring and summer, we received a boost of down from about 1,500 in a typical year, despite expecting federal funding, thanks to the excellent research projects to award 1,300 PhDs in 2009 as in recent years.
    [Show full text]
  • Tobias Lee Johnson College of Staten Island, 1S-225 [email protected] 2800 Victory Blvd
    Tobias Lee Johnson College of Staten Island, 1S-225 [email protected] 2800 Victory Blvd. http://www.math.csi.cuny.edu/~tobiasljohnson/ Staten Island, NY 10314 EMPLOYMENT AND EDUCATION Assistant Professor, College of Staten Island (CUNY) Fall 2017– NSF Postdoctoral Fellow New York University; sponsored by Gérard Ben Arous Fall 2016–Spring 2017 University of Southern California; sponsored by Larry Goldstein Fall 2014–Spring 2016 Ph.D. in Mathematics, University of Washington Fall 2008–Spring 2014 Advised by Ioana Dumitriu and Soumik Pal B.A. in Mathematics, Yale University Fall 2003–Spring 2007 RESEARCH Interests: probability theory and combinatorics, with a focus on discrete random structures and interacting particle systems; Stein’s method Papers: 20. Diffusion-limited annihilating systems and the increasing convex order, with Riti Bahl, Philip Barnet, and Matthew Junge. Submitted. arXiv:2104.12797 19. Particle density in diffusion-limited annihilating systems, with Matthew Junge, Hanbaek Lyu, and David Sivakoff. Preprint. arXiv:2005.06018 18. Continuous phase transitions on Galton–Watson trees. Submitted. arXiv:2007.13864 2020 17. Random tree recursions: which fixed points correspond to tangible sets of trees?, with Moumanti Podder and Fiona Skerman. Random Structures Algorithms, 56(3):796–837, 2020. arXiv:1808.03019 2019 16. Cover time for the frog model on trees, with Christopher Hoffman and Matthew Junge. Forum Math. Sigma, 7, e41 1–49, 2019. arXiv:1802.03428 15. Infection spread for the frog model on trees, with Christopher Hoffman and Matthew Junge. Electron. J. Probab., 24 (2019), no. 112, 1–29. arXiv:1710.05884 14. Sensitivity of the frog model to initial conditions, with Leonardo T.
    [Show full text]
  • A Personal Perspective on Mathematics Research in Industry Robert Calderbank
    comm-calderbank.qxp 4/22/98 9:13 AM Page 569 A Personal Perspective on Mathematics Research in Industry Robert Calderbank Why should a corporation such as AT&T invest more than simply the sum of the constituent de- in fundamental research in the information sci- partments. ences? At a national level, why should the fed- Scientific curiosity that bridges different dis- eral government support research in mathe- ciplines is a great spur to innovation in indus- matics? We might look at the troubles of the try, but it also has a broader educational value. National Endowment for the Arts and conclude Mathematics education is essential to an in- that beauty alone may not be enough to guar- formed understanding of many aspects of so- antee taxpayer support. Indeed the common ciety, including fairness in voting and the means wisdom is that leadership in mathematics is es- that guarantee individual privacy in the infor- sential to the nation’s competitiveness in science mation age. It is important to the long-term fu- and technology. This chain of reasoning might ture of the profession that college students leave start from, say, fundamentals of probability and mathematics departments with an appreciation lead first to Shannon’s Information Theory, then of the different ways that mathematics touches to the capacity of particular channels, and finally their everyday experience. It is not enough to to the modes of transmission that approach stop at reforming the teaching of calculus. these limits. Within industry, the challenge facing Research Within industry it is essential that an Infor- Centers in the Information Sciences (which in- mation Sciences Center have a strategic influence clude mathematics, statistics, computer science, on the development of products and services.
    [Show full text]