DOCSLIB.ORG

A Type-Driven Tensor-Based Semantics for CCG

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
A Type-Driven Tensor-Based Semantics for CCG A Type-Driven Tensor-Based Semantics for CCG Jean Maillard Stephen Clark Edward Grefenstette University of Cambridge University of Cambridge University of Oxford Computer Laboratory Computer Laboratory Department of Computer Science [email protected] [email protected] [email protected] Abstract One approach is to assume that the meanings of all words are represented by context vectors, and This paper shows how the tensor-based se- then combine those vectors using some operation, mantic framework of Coecke et al. can such as vector addition, element-wise multiplica- be seamlessly integrated with Combina- tion, or tensor product (Clark and Pulman, 2007; tory Categorial Grammar (CCG). The inte- Mitchell and Lapata, 2008). A more sophisticated gration follows from the observation that approach, which is the subject of this paper, is to tensors are linear maps, and hence can adapt the compositional process from formal se- be manipulated using the combinators of mantics (Dowty et al., 1981) and attempt to build CCG, including type-raising and compo- a distributional representation in step with the syn- sition. Given the existence of robust, tactic derivation (Coecke et al., 2010; Baroni et al., wide-coverage CCG parsers, this opens up 2013). Finally, there is a third approach using neu- the possibility of a practical, type-driven ral networks, which perhaps lies in between the compositional semantics based on distri- two described above (Socher et al., 2010; Socher butional representations. et al., 2012). Here compositional distributed rep- resentations are built using matrices operating on 1 Intoduction vectors, with all parameters learnt through a su- In this paper we show how tensor-based distribu- pervised learning procedure intended to optimise tional semantics can be seamlessly integrated with performance on some NLP task, such as syntac- Combinatory Categorial Grammar (CCG, Steed- tic parsing or sentiment analysis. The approach man (2000)), building on the theoretical discus- of Hermann and Blunsom (2013) conditions the sion in Grefenstette (2013). Tensor-based distribu- vector combination operation on the syntactic type tional semantics represents the meanings of words of the combinands, moving it a little closer to the with particular syntactic types as tensors whose se- more formal semantics-inspired approaches. mantic type matches that of the syntactic type (Co- The remainder of the Introduction gives a short ecke et al., 2010). For example, the meaning of a summary of distributional semantics. The rest of transitive verb with syntactic type (S NP)/NP is the paper introduces some mathematical notation \ a 3rd-order tensor from the tensor product space from multi-linear algebra, including Einstein nota- N S N. The seamless integration with CCG tion, and then shows how the combinatory rules of ⊗ ⊗ arises from the (somewhat trivial) observation that CCG, including type-raising and composition, can tensors are linear maps — a particular kind of be applied directly to tensor-based semantic rep- function — and hence can be manipulated using resentations. As well as describing a tensor-based CCG’s combinatory rules. semantics for CCG, a further goal of this paper is to Tensor-based semantics arises from the desire to present the compositional framework of Coecke et enhance distributional semantics with some com- al. (2010), which is based on category theory, to a positional structure, in order to make distribu- computational linguistics audience using only the tional semantics more of a complete semantic the- mathematics of multi-linear algebra. ory, and to increase its utility in NLP applica- tions. There are a number of suggestions for how 1.1 Distributional Semantics to add compositionality to a distributional seman- We assume a basic knowledge of distributional se- tics (Clarke, 2012; Pulman, 2013; Erk, 2012). mantics (Grefenstette, 1994; Schutze,¨ 1998). Re- 46 Proceedings of the EACL 2014 Workshop on Type Theory and Natural Language Semantics (TTNLS), pages 46–54, Gothenburg, Sweden, April 26-30 2014. c 2014 Association for Computational Linguistics cent inroductions to the topic include Turney and ing recent techniques from “recursive” neural net- Pantel (2010) and Clark (2014). works (Socher et al., 2010). Another possibility A potentially useful distinction for this paper, is suggested by Grefenstette et al. (2013), extend- and one not commonly made, is between distri- ing the learning technique based on linear regres- butional and distributed representations. Distri- sion from Baroni and Zamparelli (2010) in which butional representations are inherently contextual, “gold-standard” distributional representations are and rely on the frequently quoted dictum from assumed to be available for some phrases and Firth that “you shall know a word from the com- larger units. pany it keeps” (Firth, 1957; Pulman, 2013). This 2 Mathematical Preliminaries leads to the so-called distributional hypothesis that words that occur in similar contexts tend to have The tensor-based compositional process relies on similar meanings, and to various proposals for taking dot (or inner) products between vectors and how to implement this hypothesis (Curran, 2004), higher-order tensors. Dot products, and a number including alternative definitions of context; alter- of other operations on vectors and tensors, can be native weighting schemes which emphasize the conveniently written using Einstein notation (also importance of some contexts over others; alterna- referred to as the Einstein summation convention). tive similarity measures; and various dimension- In the rest of the paper we assume that the vector ality reduction schemes such as the well-known spaces are over the field of real numbers. LSA technique (Landauer and Dumais, 1997). An interesting conceptual question is whether a sim- 2.1 Einstein Notation ilar distributional hypothesis can be applied to The squared amplitude of a vector v Rn is given ∈ phrases and larger units: is it the case that sen- by: tences, for example, have similar meanings if they n v 2 = v v occur in similar contexts? Work which does ex- | | i i i=1 tend the distributional hypothesis to larger units X includes Baroni and Zamparelli (2010), Clarke Similarly, the dot product of two vectors v, w n ∈ (2012), and Baroni et al. (2013). R is given by: Distributed representations, on the other hand, n can be thought of simply as vectors (or possibly v w = viwi · higher-order tensors) of real numbers, where there Xi=1 is no a priori interpretation of the basis vectors. Denote the components of an m n real matrix Neural networks can perhaps be categorised in this × A by Aij for 1 i m and 1 j n. Then way, since the resulting vector representations are ≤ ≤ ≤ ≤ n the matrix-vector product of A and v R gives simply sequences of real numbers resulting from m ∈ a vector Av R with components: the optimisation of some training criterion on a ∈ training set (Collobert and Weston, 2008; Socher n et al., 2010). Whether these distributed represen- (Av)i = Aijvj tations can be given a contextual interpretation de- Xj=1 pends on how they are trained. We can also multiply an n m matrix A and an × One important point for this paper is that the m o matrix B to produce an n o matrix AB × × tensor-based compositional process makes no as- with components: sumptions about the interpretation of the tensors. m Hence in the remainder of the paper we make no (AB) = A B reference to how noun vectors or verb tensors, ij ik kj k=1 for example, can be acquired (which, for the case X of the higher-order tensors, is a wide open re- The previous examples are some of the most search question). However, in order to help the common operations in linear algebra, and they all reader who would prefer a more grounded dis- involve sums over repeated indices. They can be cussion, one possibility is to obtain the noun vec- simplified by introducing the Einstein summation tors using standard distributional techniques (Cur- convention: summation over the relevant range ran, 2004), and learn the higher-order tensors us- is implied on every component index that occurs 47 twice. Pairs of indices that are summed over are Thus every finite-dimensional vector is a linear known as contracted, while the remaining indices functional, and vice versa. Row and column vec- are known as free. Using this convention, the tors are examples of first-order tensors. above operations can be written as: Definition 1 (First-order tensor). Given a vector 2 space V over the field , a first-order tensor T v = vivi R | | can be defined as: v w = viwi · an element of the vector space V , • (Av)i = Aijvj, i.e. the contraction of v with a linear map T : V R, the second index of A • → a V -dimensional array of numbers Ti, for (AB)ij = AikBkj, i.e. the contraction of the • | | 1 i V . second index of A with the first of B ≤ ≤ | | These three definitions are all equivalent. Given Note how the number of free indices is always a first-order tensor described using one of these conserved between the left- and right-hand sides in definitions, it is trivial to find the two other de- these examples. For instance, while the last equa- scriptions. tion has two indices on the left and four on the right, the two extra indices on the right are con- Matrices An n m matrix A over R can be rep- × tracted. Hence counting the number of free indices resented by a two-dimensional array of real num- can be a quick way of determining what type of bers A , for 1 i n and 1 j m. ij ≤ ≤ ≤ ≤ object is given by a certain mathematical expres- Via matrix-vector multiplication, the matrix A sion in Einstein notation: no free indices means can be seen as a linear map A : Rm Rn.
Load more

Recommended publications
  • Package 'Einsum'
    Package ‘einsum’ May 15, 2021 Type Package Title Einstein Summation Version 0.1.0 Description The summation notation suggested by Einstein (1916) <doi:10.1002/andp.19163540702> is a concise mathematical notation that implicitly sums over repeated indices of n- dimensional arrays. Many ordinary matrix operations (e.g. transpose, matrix multiplication, scalar product, 'diag()', trace etc.) can be written using Einstein notation. The notation is particularly convenient for expressing operations on arrays with more than two dimensions because the respective operators ('tensor products') might not have a standardized name. License MIT + file LICENSE Encoding UTF-8 SystemRequirements C++11 Suggests testthat, covr RdMacros mathjaxr RoxygenNote 7.1.1 LinkingTo Rcpp Imports Rcpp, glue, mathjaxr R topics documented: einsum . .2 einsum_package . .3 Index 5 1 2 einsum einsum Einstein Summation Description Einstein summation is a convenient and concise notation for operations on n-dimensional arrays. Usage einsum(equation_string, ...) einsum_generator(equation_string, compile_function = TRUE) Arguments equation_string a string in Einstein notation where arrays are separated by ’,’ and the result is separated by ’->’. For example "ij,jk->ik" corresponds to a standard matrix multiplication. Whitespace inside the equation_string is ignored. Unlike the equivalent functions in Python, einsum() only supports the explicit mode. This means that the equation_string must contain ’->’. ... the arrays that are combined. All arguments are converted to arrays with
    [Show full text]
  • Notes on Manifolds
    Notes on Manifolds Justin H. Le Department of Electrical & Computer Engineering University of Nevada, Las Vegas [email protected] August 3, 2016 1 Multilinear maps A tensor T of order r can be expressed as the tensor product of r vectors: T = u1 ⊗ u2 ⊗ ::: ⊗ ur (1) We herein fix r = 3 whenever it eases exposition. Recall that a vector u 2 U can be expressed as the combination of the basis vectors of U. Transform these basis vectors with a matrix A, and if the resulting vector u0 is equivalent to uA, then the components of u are said to be covariant. If u0 = A−1u, i.e., the vector changes inversely with the change of basis, then the components of u are contravariant. By Einstein notation, we index the covariant components of a tensor in subscript and the contravariant components in superscript. Just as the components of a vector u can be indexed by an integer i (as in ui), tensor components can be indexed as Tijk. Additionally, as we can view a matrix to be a linear map M : U ! V from one finite-dimensional vector space to another, we can consider a tensor to be multilinear map T : V ∗r × V s ! R, where V s denotes the s-th-order Cartesian product of vector space V with itself and likewise for its algebraic dual space V ∗. In this sense, a tensor maps an ordered sequence of vectors to one of its (scalar) components. Just as a linear map satisfies M(a1u1 + a2u2) = a1M(u1) + a2M(u2), we call an r-th-order tensor multilinear if it satisfies T (u1; : : : ; a1v1 + a2v2; : : : ; ur) = a1T (u1; : : : ; v1; : : : ; ur) + a2T (u1; : : : ; v2; : : : ; ur); (2) for scalars a1 and a2.
    [Show full text]
  • The Mechanics of the Fermionic and Bosonic Fields: an Introduction to the Standard Model and Particle Physics
    The Mechanics of the Fermionic and Bosonic Fields: An Introduction to the Standard Model and Particle Physics Evan McCarthy Phys. 460: Seminar in Physics, Spring 2014 Aug. 27,! 2014 1.Introduction 2.The Standard Model of Particle Physics 2.1.The Standard Model Lagrangian 2.2.Gauge Invariance 3.Mechanics of the Fermionic Field 3.1.Fermi-Dirac Statistics 3.2.Fermion Spinor Field 4.Mechanics of the Bosonic Field 4.1.Spin-Statistics Theorem 4.2.Bose Einstein Statistics !5.Conclusion ! 1. Introduction While Quantum Field Theory (QFT) is a remarkably successful tool of quantum particle physics, it is not used as a strictly predictive model. Rather, it is used as a framework within which predictive models - such as the Standard Model of particle physics (SM) - may operate. The overarching success of QFT lends it the ability to mathematically unify three of the four forces of nature, namely, the strong and weak nuclear forces, and electromagnetism. Recently substantiated further by the prediction and discovery of the Higgs boson, the SM has proven to be an extraordinarily proficient predictive model for all the subatomic particles and forces. The question remains, what is to be done with gravity - the fourth force of nature? Within the framework of QFT theoreticians have predicted the existence of yet another boson called the graviton. For this reason QFT has a very attractive allure, despite its limitations. According to !1 QFT the gravitational force is attributed to the interaction between two gravitons, however when applying the equations of General Relativity (GR) the force between two gravitons becomes infinite! Results like this are nonsensical and must be resolved for the theory to stand.
    [Show full text]
  • Multilinear Algebra and Applications July 15, 2014
    Multilinear Algebra and Applications July 15, 2014. Contents Chapter 1. Introduction 1 Chapter 2. Review of Linear Algebra 5 2.1. Vector Spaces and Subspaces 5 2.2. Bases 7 2.3. The Einstein convention 10 2.3.1. Change of bases, revisited 12 2.3.2. The Kronecker delta symbol 13 2.4. Linear Transformations 14 2.4.1. Similar matrices 18 2.5. Eigenbases 19 Chapter 3. Multilinear Forms 23 3.1. Linear Forms 23 3.1.1. Definition, Examples, Dual and Dual Basis 23 3.1.2. Transformation of Linear Forms under a Change of Basis 26 3.2. Bilinear Forms 30 3.2.1. Definition, Examples and Basis 30 3.2.2. Tensor product of two linear forms on V 32 3.2.3. Transformation of Bilinear Forms under a Change of Basis 33 3.3. Multilinear forms 34 3.4. Examples 35 3.4.1. A Bilinear Form 35 3.4.2. A Trilinear Form 36 3.5. Basic Operation on Multilinear Forms 37 Chapter 4. Inner Products 39 4.1. Definitions and First Properties 39 4.1.1. Correspondence Between Inner Products and Symmetric Positive Definite Matrices 40 4.1.1.1. From Inner Products to Symmetric Positive Definite Matrices 42 4.1.1.2. From Symmetric Positive Definite Matrices to Inner Products 42 4.1.2. Orthonormal Basis 42 4.2. Reciprocal Basis 46 4.2.1. Properties of Reciprocal Bases 48 4.2.2. Change of basis from a basis to its reciprocal basis g 50 B B III IV CONTENTS 4.2.3.
    [Show full text]
  • Abstract Tensor Systems As Monoidal Categories
    Abstract Tensor Systems as Monoidal Categories Aleks Kissinger Dedicated to Joachim Lambek on the occasion of his 90th birthday October 31, 2018 Abstract The primary contribution of this paper is to give a formal, categorical treatment to Penrose’s abstract tensor notation, in the context of traced symmetric monoidal categories. To do so, we introduce a typed, sum-free version of an abstract tensor system and demonstrate the construction of its associated category. We then show that the associated category of the free abstract tensor system is in fact the free traced symmetric monoidal category on a monoidal signature. A notable consequence of this result is a simple proof for the soundness and completeness of the diagrammatic language for traced symmetric monoidal categories. 1 Introduction This paper formalises the connection between monoidal categories and the ab- stract index notation developed by Penrose in the 1970s, which has been used by physicists directly, and category theorists implicitly, via the diagrammatic languages for traced symmetric monoidal and compact closed categories. This connection is given as a representation theorem for the free traced symmet- ric monoidal category as a syntactically-defined strict monoidal category whose morphisms are equivalence classes of certain kinds of terms called Einstein ex- pressions. Representation theorems of this kind form a rich history of coherence results for monoidal categories originating in the 1960s [17, 6]. Lambek’s con- arXiv:1308.3586v1 [math.CT] 16 Aug 2013 tribution [15, 16] plays an essential role in this history, providing some of the earliest examples of syntactically-constructed free categories and most of the key ingredients in Kelly and Mac Lane’s proof of the coherence theorem for closed monoidal categories [11].
    [Show full text]
  • Quantum Chromodynamics (QCD) QCD Is the Theory That Describes the Action of the Strong Force
    Quantum chromodynamics (QCD) QCD is the theory that describes the action of the strong force. QCD was constructed in analogy to quantum electrodynamics (QED), the quantum field theory of the electromagnetic force. In QED the electromagnetic interactions of charged particles are described through the emission and subsequent absorption of massless photons (force carriers of QED); such interactions are not possible between uncharged, electrically neutral particles. By analogy with QED, quantum chromodynamics predicts the existence of gluons, which transmit the strong force between particles of matter that carry color, a strong charge. The color charge was introduced in 1964 by Greenberg to resolve spin-statistics contradictions in hadron spectroscopy. In 1965 Nambu and Han introduced the octet of gluons. In 1972, Gell-Mann and Fritzsch, coined the term quantum chromodynamics as the gauge theory of the strong interaction. In particular, they employed the general field theory developed in the 1950s by Yang and Mills, in which the carrier particles of a force can themselves radiate further carrier particles. (This is different from QED, where the photons that carry the electromagnetic force do not radiate further photons.) First, QED Lagrangian… µ ! # 1 µν LQED = ψeiγ "∂µ +ieAµ $ψe − meψeψe − Fµν F 4 µν µ ν ν µ Einstein notation: • F =∂ A −∂ A EM field tensor when an index variable µ • A four potential of the photon field appears twice in a single term, it implies summation µ •γ Dirac 4x4 matrices of that term over all the values of the index
    [Show full text]
  • About Matrices, Tensors and Various Abbrevia- Tions
    About matrices, tensors and various abbrevia- tions The objects we deal with in this course are rather complicated. We therefore use a simplifying notation where at every step as much as possible of this complexity is hidden. Basically this means that objects that are made out of many elements or components. are written without indices as much as possible. We also have several different types of indices present. The ones that show up in this course are: • Dirac-Indices a; b; c • Lorentz indices, both upper and lower µ, ν; ρ • SU(2)L indices i; j; k • SU(3)c indices α; β; γ Towards the right I have written the type of symbols used in this note to denote a particular type of index. The course contains even more, three-vector indices or various others denoting sums over types of quarks and/or leptons. An object is called scalar or singlet if it has no index of a particular type of index, a vector if it has one, a matrix if it has two and a tensor if it has two or more. A vector can also be called a column or row matrix. Examples are b with elements bi: 0 1 b1 B C B b2 C b = (b1; b2; : : : ; bn) b = (b1 b2 ··· bn) b = B . C (1) B . C @ . A bn In the first case is a vector, the second a row matrix and the last a column vector. Objects with two indices are called a matrix or sometimes a tensor and de- noted by 0 1 c11 c12 ··· c1n B C B c21 c22 ··· c2n C c = (cij) = B .
    [Show full text]
  • Tensor Algebra
    TENSOR ALGEBRA Continuum Mechanics Course (MMC) - ETSECCPB - UPC Introduction to Tensors Tensor Algebra 2 Introduction SCALAR , , ... v VECTOR vf, , ... MATRIX σε,,... ? C,... 3 Concept of Tensor A TENSOR is an algebraic entity with various components which generalizes the concepts of scalar, vector and matrix. Many physical quantities are mathematically represented as tensors. Tensors are independent of any reference system but, by need, are commonly represented in one by means of their “component matrices”. The components of a tensor will depend on the reference system chosen and will vary with it. 4 Order of a Tensor The order of a tensor is given by the number of indexes needed to specify without ambiguity a component of a tensor. a Scalar: zero dimension 3.14 1.2 v 0.3 a , a Vector: 1 dimension i 0.8 0.1 0 1.3 2nd order: 2 dimensions A, A E 02.40.5 ij rd A , A 3 order: 3 dimensions 1.3 0.5 5.8 A , A 4th order … 5 Cartesian Coordinate System Given an orthonormal basis formed by three mutually perpendicular unit vectors: eeˆˆ12,, ee ˆˆ 23 ee ˆˆ 31 Where: eeeˆˆˆ1231, 1, 1 Note that 1 if ij eeˆˆi j ij 0 if ij 6 Cylindrical Coordinate System x3 xr1 cos x(,rz , ) xr2 sin xz3 eeeˆˆˆr cosθθ 12 sin eeeˆˆˆsinθθ cos x2 12 eeˆˆz 3 x1 7 Spherical Coordinate System x3 xr1 sin cos xrxr, , 2 sin sin xr3 cos ˆˆˆˆ x2 eeeer sinθφ sin 123sin θ cos φ cos θ eeeˆˆˆ cosφφ 12sin x1 eeeeˆˆˆˆφ cosθφ sin 123cos θ cos φ sin θ 8 Indicial or (Index) Notation Tensor Algebra 9 Tensor Bases – VECTOR A vector v can be written as a unique linear combination of the three vector basis eˆ for i 1, 2, 3 .
    [Show full text]
  • Introduction to Quantum Mechanics Unit 0. Review of Linear Algebra
    Introduction to Quantum Mechanics Unit 0. Review of Linear Algebra A. Linear Space 1. Definition: A linear space is a collection S of vectors |a>, |b>, |c>, …. (usually infinite number of them) on which vector addition + is defined so that (i) (S,+) forms an Abelien group, i,e,, S is closed under + + is associate: (|a>+|b>)+|c> = |a>+(|b>+|c>) + is commutative: |a>+|b>=|b>+|a> Existence of identity: |a> + |0> = |a> for all |a> Existence of inverse: |a> + (|-a>) = |0> and also scalar (usually complex number) * is defined so that (ii) S is closed under * (iii) * and complex number multiplication is associative: α(β*|a>)=(αβ)*|a> (iv) 1|a>=|a> (and hence 0|a>=|0> and |a>=-|a>) (v) * and + is distributive: (α+β)|a>=α|a>+β|a> and α(|a>+|b>)=α|a>+α|b> 2. Basis and dimension (i) Definition of a basis: A basis is a collection of vectors B={|a1>, |a2>,|a3>,….|aN>} such that any vector in S can be written as a linear combination of a1>, |a2>,|a3>,….|aN>: a = α1 a1 + α 2 a 2 + α3 a 3 +Lα N a N where α1 ,α 2 ,Kα N are complex numbers Furthermore, ={|a1>, |a2>,|a3>,….|aN> are independent of each other, i.e. 0 = α1 a1 + α 2 a 2 + α 3 a 3 +Lα N a N if and only if α1 = α 2 = K = α N = 0 (ii) Choice of a basis is not unique. (iii) However, all basises have the same number of elements. This number (say, N) depends only on the linear space S, and it is known as the dimension of the linear space S.
    [Show full text]
  • Lost in the Tensors: Einstein's Struggles with Covariance Principles 1912-1916"
    JOHN EARMAN and CLARK GL YMOUR LOST IN THE TENSORS: EINSTEIN'S STRUGGLES WITH COVARIANCE PRINCIPLES 1912-1916" Introduction IN 1912 Einstein began to devote a major portion of his time and energy to an attempt to construct a relativistic theory of gravitation. A strong intimation of the struggle that lay ahead is contained in a letter to Arnold Sommerfeld dated October 29, 1912: At the moment I am working solely on the problem of gravitation and believe 1 will be able to overcome all difficulties with the help of a local, friendly mathemat- ician. But one thing is certain, that I have never worked so hard in my life, and that I have been injected with a great awe of mathematics, which in my naivet~ until now I only viewed as a pure luxury in its subtler forms! Compared to this problem the original theory of relativity is mere child's play.' Einstein's letter contained only a perfunctory reply to a query from Sommerfeld about the Debye-Born theory of specific heats. Obviously disappointed, Som- merfeld wrote to Hilbert: 'My letter to Einstein was in vain . Einstein is evidently so deeply mired in gravitation that he is deaf to everything else? Sommerfeld's words were more prophetic than he could possibly have known; the next three years were to see Einstein deeply mired in gravitation, sometimes seemingly hopelessly so. In large measure, Einstein's struggle resulted from his use and his misuse, his understanding and his misunderstanding of the nature and implications of covariance principles. In brief, considerations of general covariance were bound up with Einstein's motive for seeking a 'generalized' theory of relativity; mis- understandings about the meaning and implementation of this motivation threatened to wreck the search; and in the end, the desire for general covariance helped to bring Einstein back onto the track which led to what we now recognize *Present address c/o Department of Philosophy, University of Minnesota, Minneapolis, Minn, U.S.A.
    [Show full text]
  • Indicial Notation
    Appendix A Indicial Notation A.1 Indicial Notation for Vector and Matrix Operations In tensor analysis, an extensive use of indicial notation is made. Operations using Cartesian components of vectors and matrices can be expressed efficiently using this notation. A.1.1 Elements of a Matrix A set of n variables x1, x2,...,xn is denoted using indicial notation such as as xi , i = 1, 2,...,n. Let us consider now the equation that describes a plane in a three dimensional space x1, x2, x3, a1x1 + a2x2 + a3x3 = p, where ai and p are constants. This equation can be expressed alternatively as 3 ai xi = p. i=1 At this point, we can use the convention usually referred to as Einstein notation or Einstein summation convention, which implies summation over a set of indexed terms in a formula when an index appears twice in a single term along the variational rank of the indexes (1 to 3 in the above equation). Therefore, ai xi = p In the same spirit, © Springer International Publishing AG, part of Springer Nature 2018 227 E. Cueto and D. González, An Introduction to Structural Mechanics for Architects, Structural Integrity 4, https://doi.org/10.1007/978-3-319-72935-0 228 Appendix A: Indicial Notation • Vectorial scalar product: u · v = ui vi • Norm of a vector: √ √ ||u|| = u · u = ui ui • Derivative of a function: n ∂ f ∂ f df = dx = dx = f, dx ∂x i ∂x i i i i=1 i i 1 Let us define the Kronecker Delta δij as a symbol whose values can be, = δ = 1ifi j ij 0ifi = j In the same way, let us define the permutation symbol εijk as, ⎧ ⎨ 0ifi = jj= ki= k ε = , , ijk ⎩ 1ifi j k for an even number permutation −1ifi, j, k form an odd permutation A.1.
    [Show full text]
  • 1 Einstein Notation
    PH 108 : Electricity & Magnetism Einstein Notation Rwitaban Goswami 1 Einstein Notation 2.6. The gradient operator r behaves like a vector in \ some sense". For example, divergence of a curl (r:r × A~ = 0) for any A~, may suggest that it is just like A:~ B~ × C~ being zero if any two vectors are equal. Prove that r × r × F~ = r(r:F~ ) − r2F~ . To what extent does this look like the well known expansion of A~ × B~ × C~ ? Solution: We will be using Einstein summation convention. According to this convention: 1. The whole term is summed over any index which appears twice in that term 2. Any index appearing only once in the term in not summed over and must be present in both LHS and RHS Don't be confused by this. The convention only is to simply omit the P symbol. You can mentally insert the P symbol to make sense of what the summation will look like. Now let us go over some common symbols that are used extensively in Einstein's notation: • δij is the Kronecker-Delta symbol. It is defined as ( 1 if i = j δ = ij 0 if i 6= j • ijk is the Levi-Civita symbol. It is defined as 81 if ijk is a cyclic permutation of xyz, eg xyz, yzx, zxy <> ijk = −1 if ijk is an anticyclic permuation of xyz, eg xzy, zyx, yxz :>0 if in ijk any index is repeating, eg xxy, xyy etc Now, you can verify: ~ ~ A · B = AiBi ~ ~ A × B = ijkAjBk e^i wheree ^i is the unit vector in i direction In the cross product formula, when the P symbol is reinserted, this simply means X X X ijkAjBke^i i=x;y;z j=x;y;z k=x;y;z A few things to note: P • δii = 3.
    [Show full text]