DOCSLIB.ORG

Tensor Rank Decompositions Via the Pseudo-Moment Method

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
Tensor Rank Decompositions Via the Pseudo-Moment Method TENSOR RANK DECOMPOSITIONS VIA THE PSEUDO-MOMENT METHOD A Dissertation Presented to the Faculty of the Graduate School of Cornell University in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy by Jonathan Shi December 2019 © 2019 Jonathan Shi ALL RIGHTS RESERVED TENSOR RANK DECOMPOSITIONS VIA THE PSEUDO-MOMENT METHOD Jonathan Shi, Ph.D. Cornell University 2019 Over a series of four articles and an introduction, a “method of pseudo-moments” is developed which gives polynomial-time tensor rank decompositions for a variety of tensor component models. Algorithms given fall into two general classes: those falling back on convex optimization, which develop the theory of polynomial-time algorithms, as well as those constructed through spectral and matrix polynomial methods, which illustrate the possibility of realizing the aforementioned polynomial-time algorithm ideas in runtimes practical for real- life inputs. Tensor component models covered include those featuring random, worst-case, generic, or smoothed inputs, as well as cases featuring overcomplete and undercomplete rank. All models are capable of tolerating substantial noise in tensor inputs, measured in spectral norm. BIOGRAPHICAL SKETCH Jonathan Shi was raised in Seattle, Washington, and received a Bachelors of Science from the University of Washington in 2013 in the major concentrations of Computer Science, Mathematics, and Physics. He has been working at Cornell University under the guidance of Professor David Steurer and will soon join Bocconi University in Milan as a Research Fellow. iii This dissertation is dedicated to my mother and father, Richard C-J Shi and Tracey H. Luo, who made this work possible with their perseverence, and in the memory of Mary Kaitlynne Richardson, whose kindness would have driven her to greatness. iv ACKNOWLEDGEMENTS I am in gratitude for the support and mentorship of my advisor David Steurer, who oversaw my development of skill and knowledge in this thesis topic from the starting scraps I had, as well as Professor Robert Kleinberg in his role as a Director of Graduate Studies, committee member, and all-around pretty cool person. I would like to ensure the acknowledgement of those in the student community who worked (in part in student organizations) to create a welcoming, supportive, and inclusive community, as well as those who labor to keep the Computer Science Field at Cornell running. I would not have been able to complete this work without the aid of Eve Abrams, LCSW-R, Robert Mendola, MD, Clint Wattenberg, MS RD, and Edward Koppel, MD in improving my health and directing me to needed resources. I acknowledge support from a Cornell University Fellowship as well as the NSF via my advisor’s NSF CAREER Grant CCF-1350196 over the duration of my graduate program. v CONTENTS Biographical Sketch.............................. iii Dedication................................... iv Acknowledgements..............................v Contents.................................... vi List of Tables..................................x List of Figures................................. xi 1 Introduction1 1.0.1 Results stated..........................3 1.0.2 Overview of methods.....................7 1.0.3 Frontier of what is possible.................. 13 1.0.4 Organization.......................... 15 2 Spiked tensor model 17 2.1 Introduction............................... 17 2.1.1 Results.............................. 20 2.1.2 Techniques............................ 23 2.1.3 Related Work.......................... 27 2.2 Preliminaries............................... 29 2.2.1 Notation............................. 29 2.2.2 Polynomials and Matrices................... 30 2.2.3 The Sum of Squares (SoS) Algorithm............. 31 2.3 Certifying Bounds on Random Polynomials............. 32 2.4 Polynomial-Time Recovery via Sum of Squares........... 35 2.4.1 Semi-Random Tensor PCA.................. 39 2.5 Linear Time Recovery via Further Relaxation............ 40 2.5.1 The Spectral SoS Relaxation.................. 41 Í ⊗ 2.5.2 Recovery via the i Ti Ti Spectral SoS Solution...... 44 2.5.3 Nearly-Linear-Time Recovery via Tensor Unfolding and Spectral SoS........................... 49 2.5.4 Fast Recovery in the Semi-Random Model......... 52 2.5.5 Fast Recovery with Symmetric Noise............. 54 2.5.6 Numerical Simulations..................... 58 2.6 Lower Bounds.............................. 59 2.6.1 Polynomials, Vectors, Matrices, and Symmetries, Redux. 62 2.6.2 Formal Statement of the Lower Bound............ 65 2.6.3 In-depth Preliminaries for Pseudo-Expectation Symmetries 68 2.6.4 Construction of Initial Pseudo-Distributions........ 71 2.6.5 Getting to the Unit Sphere................... 75 2.6.6 Repairing Almost-Pseudo-Distributions........... 79 2.6.7 Putting Everything Together................. 80 2.7 Higher-Order Tensors......................... 87 vi 3 Spectral methods for the random overcomplete model 90 3.1 Introduction............................... 90 3.1.1 Planted Sparse Vector in Random Linear Subspace.... 92 3.1.2 Overcomplete Tensor Decomposition............ 94 3.1.3 Tensor Principal Component Analysis............ 97 3.1.4 Related Work.......................... 98 3.2 Techniques................................ 100 3.2.1 Planted Sparse Vector in Random Linear Subspace.... 104 3.2.2 Overcomplete Tensor Decomposition............ 107 3.2.3 Tensor Principal Component Analysis............ 111 3.3 Preliminaries............................... 113 3.4 Planted Sparse Vector in Random Linear Subspace......... 115 3.4.1 Algorithm Succeeds on Good Basis............. 119 3.5 Overcomplete Tensor Decomposition................. 122 3.5.1 Proof of Theorem 3.5.3..................... 125 3.5.2 Discussion of Full Algorithm................. 129 3.6 Tensor principal component analysis................. 132 3.6.1 Spiked tensor model...................... 132 3.6.2 Linear-time algorithm..................... 133 4 Polynomial lifts 136 4.1 Introduction............................... 136 4.1.1 Results for tensor decomposition............... 140 4.1.2 Applications of tensor decomposition............ 146 4.1.3 Polynomial optimization with few global optima...... 148 4.2 Techniques................................ 149 4.2.1 Rounding pseudo-distributions by matrix diagonalization 150 4.2.2 Overcomplete fourth-order tensor.............. 154 4.2.3 Random overcomplete third-order tensor.......... 157 4.3 Preliminaries............................... 158 4.3.1 Pseudo-distributions...................... 160 4.3.2 Sum of squares proofs..................... 162 4.3.3 Matrix constraints and sum-of-squares proofs....... 164 4.4 Rounding pseudo-distributions.................... 165 4.4.1 Rounding by matrix diagonalization............. 165 4.4.2 Improving accuracy of a found solution........... 169 4.5 Decomposition with sum-of-squares................. 171 4.5.1 General algorithm for tensor decomposition........ 172 4.5.2 Tensors with orthogonal components............ 175 4.5.3 Tensors with separated components............. 176 4.6 Spectral norms and tensor operations................ 179 4.6.1 Spectral norms and pseudo-distributions.......... 179 4.6.2 Spectral norm of random contraction............ 181 4.7 Decomposition of random overcomplete 3-tensors......... 185 vii 4.8 Robust decomposition of overcomplete 4-tensors.......... 189 4.8.1 Noiseless case.......................... 193 4.8.2 Noisy case............................ 195 4.8.3 Condition number under smooth analysis.......... 199 4.9 Tensor decomposition with general components.......... 201 4.9.1 Improved rounding of pseudo-distributions........ 202 4.9.2 Finding all components.................... 208 4.10 Fast orthogonal tensor decomposition without sum-of-squares.. 211 5 Overcomplete generic decomposition spectrally 216 5.1 Introduction............................... 216 5.1.1 Our Results........................... 221 5.1.2 Related works.......................... 224 5.2 Overview of algorithm......................... 230 5.3 Preliminaries............................... 235 5.4 Tools for analysis and implementation................ 238 5.4.1 Robustness and spectral perturbation............ 238 5.4.2 Efficient implementation and runtime analysis....... 239 5.5 Lifting.................................. 241 5.5.1 Algebraic identifiability argument.............. 245 5.5.2 Robustness arguments..................... 247 5.6 Rounding................................. 250 5.6.1 Recovering candidate whitened and squared components 251 5.6.2 Extracting components from the whitened squares.... 255 5.6.3 Testing candidate components................ 257 5.6.4 Putting things together..................... 259 5.6.5 Cleaning............................. 261 5.7 Combining lift and round for final algorithm............ 262 5.8 Condition number of random tensors................ 266 5.8.1 Notation............................. 272 5.8.2 Fourth Moment Identities................... 273 5.8.3 Matrix Product Identities................... 274 5.8.4 Naive Spectral Norm Estimate................ 277 5.8.5 Off-Diagonal Second Moment Estimates........... 278 5.8.6 Matrix Decoupling....................... 280 5.8.7 Putting It Together....................... 281 5.8.8 Omitted Proofs......................... 283 Bibliography 286 A Spiked tensor model 294 A.1 Pseudo-Distribution Facts....................... 294 A.2 Concentration bounds......................... 297 A.2.1 Elementary Random Matrix Review............. 297 viii Í ⊗ A.2.2 Concentration
Load more

Recommended publications
  • Matrix-Valued Moment Problems
    Matrix-valued moment problems A Thesis Submitted to the Faculty of Drexel University by David P. Kimsey in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Mathematics June 2011 c Copyright 2011 David P. Kimsey. All Rights Reserved. ii Acknowledgments I wish to thank my advisor Professor Hugo J. Woerdeman for being a source of mathematical and professional inspiration. Woerdeman's encouragement and support have been extremely beneficial during my undergraduate and graduate career. Next, I would like to thank Professor Robert P. Boyer for providing a very important men- toring role which guided me toward a career in mathematics. I owe many thanks to L. Richard Duffy for providing encouragement and discussions on fundamental topics. Professor Dmitry Kaliuzhnyi-Verbovetskyi participated in many helpful discussions which helped shape my understanding of several mathematical topics of interest to me. In addition, Professor Anatolii Grinshpan participated in many useful discussions gave lots of encouragement. I wish to thank my Ph.D. defense committee members: Professors Boyer, Lawrence A. Fialkow, Grinshpan, Pawel Hitczenko, and Woerde- man for pointing out any gaps in my understanding as well as typographical mistakes in my thesis. Finally, I wish to acknowledge funding for my research assistantship which was provided by the National Science Foundation, via the grant DMS-0901628. iii Table of Contents Abstract ................................................................................ iv 1. INTRODUCTION ................................................................
    [Show full text]
  • Geometry of the Pfaff Lattices 3
    GEOMETRY OF THE PFAFF LATTICES YUJI KODAMA∗ AND VIRGIL U. PIERCE∗∗ Abstract. The (semi-infinite) Pfaff lattice was introduced by Adler and van Moerbeke [2] to describe the partition functions for the random matrix models of GOE and GSE type. The partition functions of those matrix models are given by the Pfaffians of certain skew-symmetric matrices called the moment matrices, and they are the τ-functions of the Pfaff lattice. In this paper, we study a finite version of the Pfaff lattice equation as a Hamiltonian system. In particular, we prove the complete integrability in the sense of Arnold-Liouville, and using a moment map, we describe the real isospectral varieties of the Pfaff lattice. The image of the moment map is a convex polytope whose vertices are identified as the fixed points of the flow generated by the Pfaff lattice. Contents 1. Introduction 2 1.1. Lie algebra splitting related to SR-factorization 2 1.2. Hamiltonian structure 3 1.3. Outline of the paper 5 2. Integrability 6 2.1. The integrals Fr,k(L) 6 2.2. The angle variables conjugate to Fr,k(L) 10 2.3. Extended Jacobian 14 3. Matrix factorization and the τ-functions 16 3.1. Moment matrix and the τ-functions 17 3.2. Foliation of the phase space by Fr,k(L) 21 3.3. Examples from the matrix models 22 4. Real solutions 23 arXiv:0705.0510v1 [nlin.SI] 3 May 2007 4.1. Skew-orthogonal polynomials 23 4.2. Fixed points of the Pfaff flows 26 4.3.
    [Show full text]
  • Pre-Publication Accepted Manuscript
    Peter Forrester, Shi-Hao Li Classical discrete symplectic ensembles on the linear and exponential lattice: skew orthogonal polynomials and correlation functions Transactions of the American Mathematical Society DOI: 10.1090/tran/7957 Accepted Manuscript This is a preliminary PDF of the author-produced manuscript that has been peer-reviewed and accepted for publication. It has not been copyedited, proofread, or finalized by AMS Production staff. Once the accepted manuscript has been copyedited, proofread, and finalized by AMS Production staff, the article will be published in electronic form as a \Recently Published Article" before being placed in an issue. That electronically published article will become the Version of Record. This preliminary version is available to AMS members prior to publication of the Version of Record, and in limited cases it is also made accessible to everyone one year after the publication date of the Version of Record. The Version of Record is accessible to everyone five years after publication in an issue. CLASSICAL DISCRETE SYMPLECTIC ENSEMBLES ON THE LINEAR AND EXPONENTIAL LATTICE: SKEW ORTHOGONAL POLYNOMIALS AND CORRELATION FUNCTIONS PETER J. FORRESTER AND SHI-HAO LI Abstract. The eigenvalue probability density function for symplectic invariant random matrix ensembles can be generalised to discrete settings involving either a linear or exponential lattice. The corresponding correlation functions can be expressed in terms of certain discrete, and q, skew orthogonal polynomials respectively. We give a theory of both of these classes of polynomials, and the correlation kernels determining the correlation functions, in the cases that the weights for the corresponding discrete unitary ensembles are classical.
    [Show full text]
  • Representational Models: a Common Framework for Understanding Encoding, Pattern-Component, and Representational-Similarity Analysis
    Representational models: A common framework for understanding encoding, pattern-component, and representational-similarity analysis Jörn Diedrichsen1 & Nikolaus Kriegeskorte2 1. Brain and Mind Institute, Western University, Canada 2. Cognitive and Brain Sciences Unit, Cambridge University, UK Abstract Representational models explain how activity patterns in populations of neurons (or, more generally, in multivariate brain activity measurements) relate to sensory stimuli, motor actions, or cognitive processes. In an experimental context, representational models can be defined as probabilistic hypotheses about what profiles of activations across conditions are likely to be observed. We describe three methods to test such models – encoding models, pattern component modeling (PCM), and representational similarity analysis (RSA). We show that these methods are closely related in that they all evaluate the statistical second moment of the activity profile distribution. Using simulated data from three different fMRI experiments, we compare the power of the approaches to adjudicate between competing representational models. PCM implements a likelihood-ratio test and therefore constitutes the most powerful test if its assumptions hold. However, the other two approaches – when conducted appropriately – can perform similarly. In encoding approaches, the linear model needs to be appropriately regularized, which imposes a prior on the activity profiles. Without such a prior, encoding approaches do not test well-defined representational models. In RSA, the unequal variances and dependencies of the distance measures can be taken into account using a multi-normal approximation to the sampling distribution of the distance measures, so as to enable optimal inference. Each of the three techniques renders different information explicit (e.g. single response tuning in encoding models and population representational dissimilarity in RSA) and has specific advantages in terms of computational demands, ease of use, and extensibility.
    [Show full text]
  • A Panorama of Positivity
    A PANORAMA OF POSITIVITY ALEXANDER BELTON, DOMINIQUE GUILLOT, APOORVA KHARE, AND MIHAI PUTINAR Abstract. This survey contains a selection of topics unified by the concept of positive semi-definiteness (of matrices or kernels), reflecting natural constraints imposed on discrete data (graphs or networks) or continuous objects (probability or mass distributions). We put empha- sis on entrywise operations which preserve positivity, in a variety of guises. Techniques from harmonic analysis, function theory, operator theory, statistics, combinatorics, and group representations are invoked. Some partially forgotten classical roots in metric geometry and distance transforms are presented with comments and full bibliographical refer- ences. Modern applications to high-dimensional covariance estimation and regularization are included. Contents 1. Introduction 2 2. From metric geometry to matrix positivity 4 2.1. Distance geometry 4 2.2. Spherical distance geometry 6 2.3. Distance transforms 6 2.4. Altering Euclidean distance 8 2.5. Positive definite functions on homogeneous spaces 11 2.6. Connections to harmonic analysis 15 3. Entrywise functions preserving positivity in all dimensions 18 Date: November 13, 2019. 2010 Mathematics Subject Classification. 15-02, 26-02, 15B48, 51F99, 15B05, 05E05, arXiv:1812.05482v3 [math.CA] 12 Nov 2019 44A60, 15A24, 15A15, 15A45, 15A83, 47B35, 05C50, 30E05, 62J10. Key words and phrases. metric geometry, positive semidefinite matrix, Toeplitz ma- trix, Hankel matrix, positive definite function, completely monotone functions, absolutely monotonic functions, entrywise calculus, generalized Vandermonde matrix, Schur polyno- mials, symmetric function identities, totally positive matrices, totally non-negative matri- ces, totally positive completion problem, sample covariance, covariance estimation, hard / soft thresholding, sparsity pattern, critical exponent of a graph, chordal graph, Loewner monotonicity, convexity, and super-additivity.
    [Show full text]
  • New Publications Offered by The
    New Publications Offered by the AMS To subscribe to email notification of new AMS publications, please go to http://www.ams.org/bookstore-email. Algebra and Algebraic Introduction Geometry to Orthogonal, Symplectic and Unitary The Schrödinger Representations of Model for the Minimal Finite Groups Representation of the Carl R. Riehm, McMaster Indefinite Orthogonal University, Hamilton, ON, Group O(p; q) Canada, and The Fields Institute, Toronto, ON, Canada Toshiyuki Kobayashi, University of Tokyo, Japan, and Gen Mano, Orthogonal, symplectic and unitary representations of finite groups lie at the crossroads of two more traditional subjects of PricewaterhouseCoopers Aarata, mathematics—linear representations of finite groups, and the Tokyo, Japan theory of quadratic, skew symmetric and Hermitian forms—and thus inherit some of the characteristics of both. Contents: Introduction; Two models of the minimal representation of O(p; q); K-finite eigenvectors in the Schrödinger model L2(C); This book is written as an introduction to the subject and not as an Radial part of the inversion; Main theorem; Bessel distributions; encyclopaedic reference text. The principal goal is an exposition of Appendix: special functions; Bibliography; List of Symbols; Index. the known results on the equivalence theory, and related matters such as the Witt and Witt-Grothendieck groups, over the “classical” Memoirs of the American Mathematical Society, Volume 213, fields—algebraically closed, real closed, finite, local and global. A Number 1000 detailed exposition of the background material needed is given in August 2011, 132 pages, Softcover, ISBN: 978-0-8218-4757-2, the first chapter. 2010 Mathematics Subject Classification: 22E30; 22E46, 43A80, It was A.
    [Show full text]
  • RANK SHIFT CONDITIONS and REDUCTIONS of 2D-TODA THEORY 3 Expressed in Terms of the Cauchy Two-Matrix Model
    RANK SHIFT CONDITIONS AND REDUCTIONS OF 2D-TODA THEORY SHI-HAO LI AND GUO-FU YU Abstract. This paper focuses on different reductions of 2-dimensional (2d-)Toda hierarchy. Symmetric and skew symmetric moment matrices are firstly considered, resulting in the differ- ential relations between symmetric/skew symmetric tau functions and 2d-Toda’s tau functions respectively. Furthermore, motivated by the Cauchy two-matrix model and Bures ensemble from random matrix theory, we study the rank one shift condition in symmetric case and rank two shift condition in skew symmetric case, from which the C-Toda hierarchy and B-Toda hierarchy are found, together with their special Lax matrices and integrable structures. 1. Introduction The studies in random matrix theory and classic integrable systems promote the development of both communities, and the orthogonal polynomials play an important role to connect these two totally different subjects. For example, in [17, §2], it is shown that these polynomials can lead to an iso-spectral flow of the Toda lattice, while the normalisation factors of the orthogonal polynomials can be expressed in terms of the partition function of the unitary invariant Hermitian matrix model. Later on, several novel random matrix models have been found in the course of doing analysis of the classical integrable systems. An example is the appearance of the Cauchy two-matrix model. This model was proposed in the studies of the peakon solutions of the Degasperis-Procesi equation and its related Hermite-Padé approximation problem [34]. To some extent, the relation between random matrix and integrable system can be described in terms of the partition function and tau functions.
    [Show full text]
  • Sums of Squares, Moment Matrices and Optimization Over Polynomials
    SUMS OF SQUARES, MOMENT MATRICES AND OPTIMIZATION OVER POLYNOMIALS MONIQUE LAURENT∗ May 8, 2008 Abstract. We consider the problem of minimizing a polynomial over a semialge- braic set defined by polynomial equations and inequalities, which is NP-hard in general. Hierarchies of semidefinite relaxations have been proposed in the literature, involving positive semidefinite moment matrices and the dual theory of sums of squares of poly- nomials. We present these hierarchies of approximations and their main properties: asymptotic/finite convergence, optimality certificate, and extraction of global optimum solutions. We review the mathematical tools underlying these properties, in particular, some sums of squares representation results for positive polynomials, some results about moment matrices (in particular, of Curto and Fialkow), and the algebraic eigenvalue method for solving zero-dimensional systems of polynomial equations. We try whenever possible to provide detailed proofs and background. Key words. positive polynomial, sum of squares of polynomials, moment problem, polynomial optimization, semidefinite programming. AMS(MOS) subject classifications. 13P10, 13J25, 13J30, 14P10, 15A99, 44A60, 90C22, 90C30. Contents 1 Introduction...............................158 1.1 The polynomial optimization problem . 159 1.2 Thescopeofthispaper . .161 1.3 Preliminaries on polynomials and semidefinite programs . 162 1.4 Contentsofthepaper . .166 2 Algebraic preliminaries . 166 2.1 Polynomial ideals and varieties . 166 2.2 The quotient algebra R[x]/I .................169 2.3 Gr¨obner bases and standard monomial bases . 171 2.4 Solving systems of polynomial equations . 173 3 Positive polynomials and sums of squares . 178 3.1 Somebasicfacts ........................178 3.2 Sums of squares and positive polynomials: Hilbert’s result . 178 3.3 Recognizing sums of squares of polynomials .
    [Show full text]
  • Arxiv:2006.16213V4 [Math.FA] 9 Jan 2021 41,4B4(Secondary)
    TOTALLY POSITIVE KERNELS, POLYA´ FREQUENCY FUNCTIONS, AND THEIR TRANSFORMS ALEXANDER BELTON, DOMINIQUE GUILLOT, APOORVA KHARE, AND MIHAI PUTINAR Abstract. The composition operators preserving total non-negativity and total pos- itivity for various classes of kernels are classified, following three themes. Letting a function act by post composition on kernels with arbitrary domains, it is shown that such a composition operator maps the set of totally non-negative kernels to itself if and only if the function is constant or linear, or just linear if it preserves total positiv- ity. Symmetric kernels are also discussed, with a similar outcome. These classification results are a byproduct of two matrix-completion results and the second theme: an extension of A.M. Whitney’s density theorem from finite domains to subsets of the real line. This extension is derived via a discrete convolution with modulated Gauss- ian kernels. The third theme consists of analyzing, with tools from harmonic analysis, the preservers of several families of totally non-negative and totally positive kernels with additional structure: continuous Hankel kernels on an interval, P´olya frequency functions, and P´olya frequency sequences. The rigid structure of post-composition transforms of totally positive kernels acting on infinite sets is obtained by combining several specialized situations settled in our present and earlier works. Contents 1. Introduction and main results 2 2. Preliminaries and overview 7 3. Total non-negativity preservers 9 4. Total-positivity preservers. I. Semi-finite domains 14 5. Total-positivity preservers are continuous 22 6. Extensions of Whitney’s approximation theorem 24 7. Totally non-negative and totally positive Hankel kernels 30 arXiv:2006.16213v4 [math.FA] 9 Jan 2021 8.
    [Show full text]
  • Lecture 1: Bivariate Linear Model
    Covariance Structure Analysis (LISREL) Professor Ross L. Matsueda Lecture Notes Do not copy, quote, or cite without permission LECTURE 1: MOMENTS AND PARAMETERS IN A BIVARIATE LINEAR STRUCTURAL MODEL I PRELIMINARIES: STRUCTURAL EQUATION MODELS. II. POPULATION MOMENTS AND PARAMETERS. III. CORRELATIONS AND STANDARDIZED COEFFICIENTS. IV. ESTIMATION AND TESTING. I. PRELIMINARIES: STRUCTURAL EQUATION MODELS This course is about the use of structural equation models for examining social phenomena. There are four basic principles involved: 1. Assume we can characterize a social phenomenon by a set of random variables. A random variable is one whose values or outcomes are governed by a probability distribution. A set of random variables have joint outcomes, which are governed by a joint probability distribution. We characterize this joint distribution by observable moments (means, variances, and covariances). 2. We assume that some underlying causal structure or model generates observed moments (covariances) The underlying structural model represents our (parsimonious) theory of the phenomenon. The model generates the data (joint distribution of random variables), which we characterize by observed moments. The parameters of the model are assumed to be invariant: they are somewhat stable over time and only change when there is true social change. The causal structure can be expressed in linear structural equations and path diagrams. Hence the terms, "covariance structure analysis," "path analysis" and "structural equation models." 3. Given observed population moments, we could compute population parameters of our structural model (assuming the model is identified). 4. However, we lack access to population moments, and therefore, we cannot compute population parameters. Instead, we have access only to sample moments.
    [Show full text]
  • A Geometric and Topological Viewpoint
    IOP Journal of Physics A: Mathematical and Theoretical J. Phys. A: Math. Theor. Journal of Physics A: Mathematical and Theoretical J. Phys. A: Math. Theor. 51 (2018) 353001 (39pp) https://doi.org/10.1088/1751-8121/aacecf 51 Topical Review 2018 © 2018 IOP Publishing Ltd Fifty years of the fnite nonperiodic Toda lattice: a geometric and topological JPHAC5 viewpoint 353001 Yuji Kodama1,3 and Barbara A Shipman2 Y Kodama and B A Shipman 1 Department of Mathematics, Ohio State University, Columbus, OH 43210, United States of America 2 Department of Mathematics, The University of Texas at Arlington, Arlington, TX Fifty years of the fnite nonperiodic Toda lattice TX 76010, United States of America E-mail: [email protected] and [email protected] Printed in the UK Received 26 January 2018, revised 20 June 2018 JPA Accepted for publication 25 June 2018 Published 23 July 2018 10.1088/1751-8121/aacecf Abstract In 1967, Japanese physicist Morikazu Toda published a pair of seminal papers in the Journal of the Physical Society of Japan that exhibited soliton solutions to a chain of particles with nonlinear interactions between nearest neighbors. In the ffty years that followed, Toda’s system of particles has been generalized in different directions, each with its own analytic, geometric, and topological 1751-8121 characteristics. These are known collectively as the Toda lattice. This survey recounts and compares the various versions of the fnite nonperiodic Toda lattice from the perspective of their geometry and topology. In particular, we highlight the polytope structure of the solution spaces as viewed through the moment map, and we explain the connection between the real indefnite Toda fows and the integral cohomology of real fag varieties.
    [Show full text]
  • Sums of Squares, Moment Matrices and Optimization Over Polynomials
    SUMS OF SQUARES, MOMENT MATRICES AND OPTIMIZATION OVER POLYNOMIALS MONIQUE LAURENT∗ Updated version: February 6, 2010 Abstract. We consider the problem of minimizing a polynomial over a semialge- braic set defined by polynomial equations and inequalities, which is NP-hard in general. Hierarchies of semidefinite relaxations have been proposed in the literature, involving positive semidefinite moment matrices and the dual theory of sums of squares of poly- nomials. We present these hierarchies of approximations and their main properties: asymptotic/finite convergence, optimality certificate, and extraction of global optimum solutions. We review the mathematical tools underlying these properties, in particular, some sums of squares representation results for positive polynomials, some results about moment matrices (in particular, of Curto and Fialkow), and the algebraic eigenvalue method for solving zero-dimensional systems of polynomial equations. We try whenever possible to provide detailed proofs and background. Key words. positive polynomial, sum of squares of polynomials, moment problem, polynomial optimization, semidefinite programming AMS(MOS) subject classifications. 13P10, 13J25, 13J30, 14P10, 15A99, 44A60, 90C22, 90C30 Contents 1 Introduction............................... 3 1.1 The polynomial optimization problem . 4 1.2 Thescopeofthispaper .................... 6 1.3 Preliminaries on polynomials and semidefinite programs . 7 1.3.1 Polynomials ..................... 7 1.3.2 Positive semidefinite matrices . 8 1.3.3 Flat extensions of matrices . 9 1.3.4 Semidefinite programs . 9 1.4 Contentsofthepaper ..................... 11 2 Algebraic preliminaries . 11 2.1 Polynomial ideals and varieties . 11 2.2 The quotient algebra R[x]/I ................. 14 2.3 Gr¨obner bases and standard monomial bases . 16 2.4 Solving systems of polynomial equations .
    [Show full text]