DOCSLIB.ORG

An Efficient Exact-PGA Algorithm for Constant Curvature Manifolds

Total Page：16

File Type：pdf, Size：1020Kb

Download full-text PDF Read full-text
An Efficient Exact-PGA Algorithm for Constant Curvature Manifolds An efficient Exact-PGA algorithm for constant curvature manifolds Rudrasis Chakraborty1, Dohyung Seo2, and Baba C. Vemuri1 1Department of CISE, University of Florida, FL 32611, USA 2U-Systems, A GE Healthcare Company, CA, USA 1 2 {rudrasis, vemuri}@cise.ufl.edu {dhseo.118}@gmail.com Abstract 1. Introduction Principal Component Analysis (PCA) is a widely used dimensionality reduction technique in Science and Engineering. PCA however requires the input data to lie in a vector space. With the advent of new technolo- Manifold-valued datasets are widely encountered in gies and wide spread use of sophisticated feature ex- many computer vision tasks. A non-linear analog of traction methods, manifold-valued data have become the PCA algorithm, called the Principal Geodesic Anal- ubiquitous in many fields including but not limited ysis (PGA) algorithm suited for data lying on Rieman- to, Computer Vision, Medical Imaging and Machine nian manifolds was reported in literature a decade ago. Learning. A nonlinear version of PCA, called the Prin- Since the objective function in the PGA algorithm is cipal Geodesic Analysis (PGA), for data lying on Rie- highly non-linear and hard to solve efficiently in gen- mannian manifolds was introduced in [8]. eral, researchers have proposed a linear approximation. Since the objective function of PGA is highly non- Though this linear approximation is easy to compute, it linear and hard to solve in general, researchers pro- lacks accuracy especially when the data exhibits a large posed a linearized version of the PGA [8]. Though this variance. Recently, an alternative called the exact PGA linearized PGA, hereafter referred to as PGA, is com- was proposed which tries to solve the optimization with- putationally efficient, it lacks accuracy for data with out any linearization. For general Riemannian mani- large spread/variance. In order to solve the objective folds, though it yields a better accuracy than the orig- function exactly, Sommer et al., [25] proposed to solve inal (linearized) PGA, for data that exhibit large vari- the original objective function (not the approximation) ance, the optimization is not computationally efficient. and called it exact PGA . While exact PGA attempts In this paper, we propose an efficient exact PGA al- to solve this complex nonlinear optimization problem, gorithm for constant curvature Riemannian manifolds it is however computationally inefficient. Though it (CCM-EPGA). The CCM-EPGA algorithm differs sig- is not possible to efficiently and accurately solve this nificantly from existing PGA algorithms in two aspects, optimization problem for a general manifold, however, (i) the distance between a given manifold-valued data for manifolds with constant sectional curvature, we for- point and the principal submanifold is computed an- mulate an efficient and exact PGA algorithm, dubbed alytically and thus no optimization is required as in CCM-EPGA. It is well known in geometry, by virtue the existing methods. (ii) Unlike the existing PGA al- of the Killing-Hopf theorem [4], that any non-zero con- gorithms, the descent into codimension-1 submanifolds stant curvature manifold is isomorphic to either the hy- does not require any optimization but is accomplished persphere (SN ) or the hyperbolic space (HN ), hence in through the use of the Rimeannian inverse Exponential this work, we present the CCM-EPGA formulation for map and the parallel transport operations. We present (SN ) and (HN ). Our formulation has several applica- theoretical and experimental results for constant cur- tions to Computer Vision and Statistics including di- vature Riemannian manifolds depicting favorable per- rectional data [21] and color spaces [19]. Several other formance of the CCM-EPGA algorithm compared to applications of hyperbolic geometry are, shape analy- existing PGA algorithms. We also present data recon- sis [30], Electrical Impedance Tomography, Geoscience struction from the principal components which has not Imaging [28], Brain Morphometry [29], Catadiaoptric been reported in literature in this setting. Vision [3] etc. In order to depict the effectiveness of our proposed 13976 CCM-EPGA algorithm, we use the average projection to an efficient and novel algorithm for exact PGA on error as defined in [25]. We also report the computa- constant curvature manifolds (CCM-EPGA). tional time comparison of the CCM-EPGA with the Let M be a Riemannian manifold. Let us sup- PGA [8] and the exact PGA [25] algorithms respec- pose we are given a dataset, X = x , , x , where { 1 ··· n} tively. Several variants of the PGA exist in literature xj M. Let us assume that the finite sample Fr´echet and we briefly mention a few here. In [23], authors mean∈ [9] of the data set exists and be denoted by µ. Let computed the principal geodesics (without approxima- Vk be the space spanned by mutually orthogonal vec- tion) only for a special Lie group, SO(3). Geodesic tors (principal directions) v1, , vk , vj TµM, j. th { ··· } ∈ ∀ PCA (GPCA) [14, 13] solves a different optimization Let Sk be the k geodesic subspace of TµM, i.e., function namely, optimizing the projection error along Sk = Expµ(Vk), where Exp is the Riemannian expo- the geodesics. Authors in [13], minimize the projection nential map (see [4] for definition). Then, the principal error instead of maximizing variance in geodesic sub- directions, vi are defined recursively by spaces (defined later in the paper). GPCA does not n use a linear approximation, but it is restricted to man- 1 2 vi = arg max d (µ, ΠSi (xj)) (1) ifolds where a closed form expression for the geodesics v v ⊥ n =1, Vi−1 j=1 exists. More recently, a probabilistic version of PGA k k ∈ X Si = Expµ(spanVi 1, vi) (2) called PPGA was presented in [31], which is a nonlinear − version of PPCA [27]. None of these methods attempt where d(x, y) is the geodesic distance between x M to compute the solution to the exact PGA problem and y M,Π (x) is the point in S closest to x ∈M. defined in [25]. Another recent work in [11], reports S The PGA∈ algorithm on M is summarized in Alg.∈1. a non-linear generalization of PGA, namely the prin- cipal geodesic curves, and argues about its usefulness Algorithm 1 The PGA algorithm on manifold M over PGA. 1: The rest of the paper is organized as follows. In Given a data set X = x1, , xn M, and 1 L dim(M) { ··· } ∈ ≤ Section 2, we present the formulation of PGA. We also ≤ discuss the details of the linearized version of PGA [8] 2: Compute the FM, µ, of X [1] and exact PGA [25]. Our formulation of CCM-EPGA 3: Set k 1 4: ←0 0 is presented in Section 2. Experimental results for the Set x¯1, , x¯n x1, , xn 5: while{ k··· L do} ← { ··· } CCM-EPGA algorithm along with comparisons to ex- ≤ n 1 act PGA and PGA are presented in Section 3. In ad- 6: 2 k 1 Solve vk = arg max d (µ, ΠSk (¯xj − )) dition to synthetic data experiments, we present the v v v ⊥ n =1, TµM, Vk−1 j=1 k k ∈ ∈ X comparative performance of CCM-EPGA on two real as in Eq. (1). data applications. In Section 4, we present the for- 7: k 1 k 1 Project x¯1− , , x¯n− to a k co-dimension mulation for the reconstruction of data from principal one submanifold{ Z···of M, which} is orthogonal to directions and components in this nonlinear setting. the current geodesic subspace. Finally, in section 5, we draw conclusions. 8: k k Set the projected points to x¯1 , , x¯n 9: k k + 1 { ··· } 2. Principal Geodesic Analysis 10: end while← Principal Component Analysis (PCA) [17] is a well known and widely used statistical method for dimen- 2.1. PGA and exact PGA sionality reduction. Given a vector valued dataset, it returns a sequence of linear subspaces that maximize In Alg. 1 (lines 6 7), as the projection operator Π the variance of the projected data. The kth subspace is hard to compute,− hence a common alternative is to is spanned by the principal vectors v1, v2, , vk locally linearize the manifold. This approach [8] maps which are mutually orthogonal. PCA{ is well··· suited} all data points on to the tangent space at µ, and as for vector-valued data sets but not for manifold-valued the tangent plane is a vector space, one can use the inputs. A decade ago, the nonlinear version called the PCA to compute the principal directions. This simple Principal Geodesic Analysis (PGA) was developed to scheme is an approximation to the PGA and naturally cope with manifold-valued inputs [8]. In this section, raises the following question: Is it possible to do PGA first, we briefly describe this PGA algorithm, then, we (solve Eq. (1)) without any linearization? The answer show the key modification performed in [25] to arrive is yes. But, computation of the projection operator, at what they termed as the exact PGA algorithm. We ΠS(x), i.e., the closest point to x in S is computation- then motivate and present our approach which leads ally expensive. In [25], Sommer et al. give an alterna- 3977 1 ¯ θ ¯ tive formulation for the PGA by minimizing the aver- by, Expψ− (ψ) = sin(θ) (ψ ψ cos(θ)) where, θ = 2 − age squared reconstruction error, i.e., d (xj, ΠSi (xj)) d(ψ, ψ¯). 2 instead of d (µ, ΠSi (xj)) in eqns. (1). They use an op- timization scheme to compute this projection. Further, 2.2.2 Basic Riemannian Geometry of HN they termed their algorithm, exact PGA, as it does not require any linearization. However, their optimization The hyperbolic N-dimensional manifold can be embed- scheme is in general computationally expensive and for ded in RN+1 using any of three different models.
Load more

Recommended publications
  • Chapter 13 Curvature in Riemannian Manifolds
    Chapter 13 Curvature in Riemannian Manifolds 13.1 The Curvature Tensor If (M, , )isaRiemannianmanifoldand is a connection on M (that is, a connection on TM−), we− saw in Section 11.2 (Proposition 11.8)∇ that the curvature induced by is given by ∇ R(X, Y )= , ∇X ◦∇Y −∇Y ◦∇X −∇[X,Y ] for all X, Y X(M), with R(X, Y ) Γ( om(TM,TM)) = Hom (Γ(TM), Γ(TM)). ∈ ∈ H ∼ C∞(M) Since sections of the tangent bundle are vector fields (Γ(TM)=X(M)), R defines a map R: X(M) X(M) X(M) X(M), × × −→ and, as we observed just after stating Proposition 11.8, R(X, Y )Z is C∞(M)-linear in X, Y, Z and skew-symmetric in X and Y .ItfollowsthatR defines a (1, 3)-tensor, also denoted R, with R : T M T M T M T M. p p × p × p −→ p Experience shows that it is useful to consider the (0, 4)-tensor, also denoted R,givenby R (x, y, z, w)= R (x, y)z,w p p p as well as the expression R(x, y, y, x), which, for an orthonormal pair, of vectors (x, y), is known as the sectional curvature, K(x, y). This last expression brings up a dilemma regarding the choice for the sign of R. With our present choice, the sectional curvature, K(x, y), is given by K(x, y)=R(x, y, y, x)but many authors define K as K(x, y)=R(x, y, x, y). Since R(x, y)isskew-symmetricinx, y, the latter choice corresponds to using R(x, y)insteadofR(x, y), that is, to define R(X, Y ) by − R(X, Y )= + .
    [Show full text]
  • 3. Sectional Curvature of Lorentzian Manifolds. 1
    3. SECTIONAL CURVATURE OF LORENTZIAN MANIFOLDS. 1. Sectional curvature, the Jacobi equation and \tidal stresses". The (3,1) Riemann curvature tensor has the same definition in the rie- mannian and Lorentzian cases: R(X; Y )Z = rX rY Z − rY rX Z − r[X;Y ]Z: If f(t; s) is a parametrized 2-surface in M (immersion) and W (t; s) is a vector field on M along f, we have the Ricci formula: D DW D DW − = R(@ f; @ f)W: @t @s @s @t t s For a variation f(t; s) = γs(t) of a geodesic γ(t) (with variational vector field J(t) = @sfjs=0 along γ(t)) this leads to the Jacobi equation for J: D2J + R(J; γ_ )_γ = 0: dt2 ? The Jacobi operator is the self-adjoint operator on (γ _ ) : Rp[v] = Rp(v; γ_ )_γ. When γ is a timelike geodesic (the worldline of a free-falling massive particle) the physical interpretation of J is the relative displacement (space- like) vector of a neighboring free-falling particle, while the second covariant 00 derivative J represents its relative acceleration. The Jacobi operator Rp gives the \tidal stresses" in terms of the position vector J. In the Lorentzian case, the sectional curvature is defined only for non- degenerate two-planes Π ⊂ TpM. Definition. Let Π = spanfX; Y g be a non-degenerate two-dimensional subspace of TpM. The sectional curvature σXY = σΠ is the real number σ defined by: hR(X; Y )Y; Xi = σhX ^ Y; X ^ Y i: Remark: by a result of J.
    [Show full text]
  • Hyperbolic Geometry
    Flavors of Geometry MSRI Publications Volume 31,1997 Hyperbolic Geometry JAMES W. CANNON, WILLIAM J. FLOYD, RICHARD KENYON, AND WALTER R. PARRY Contents 1. Introduction 59 2. The Origins of Hyperbolic Geometry 60 3. Why Call it Hyperbolic Geometry? 63 4. Understanding the One-Dimensional Case 65 5. Generalizing to Higher Dimensions 67 6. Rudiments of Riemannian Geometry 68 7. Five Models of Hyperbolic Space 69 8. Stereographic Projection 72 9. Geodesics 77 10. Isometries and Distances in the Hyperboloid Model 80 11. The Space at Infinity 84 12. The Geometric Classification of Isometries 84 13. Curious Facts about Hyperbolic Space 86 14. The Sixth Model 95 15. Why Study Hyperbolic Geometry? 98 16. When Does a Manifold Have a Hyperbolic Structure? 103 17. How to Get Analytic Coordinates at Infinity? 106 References 108 Index 110 1. Introduction Hyperbolic geometry was created in the first half of the nineteenth century in the midst of attempts to understand Euclid’s axiomatic basis for geometry. It is one type of non-Euclidean geometry, that is, a geometry that discards one of Euclid’s axioms. Einstein and Minkowski found in non-Euclidean geometry a This work was supported in part by The Geometry Center, University of Minnesota, an STC funded by NSF, DOE, and Minnesota Technology, Inc., by the Mathematical Sciences Research Institute, and by NSF research grants. 59 60 J. W. CANNON, W. J. FLOYD, R. KENYON, AND W. R. PARRY geometric basis for the understanding of physical time and space. In the early part of the twentieth century every serious student of mathematics and physics studied non-Euclidean geometry.
    [Show full text]
  • Global Properties of Asymptotically De Sitter and Anti De Sitter Spacetimes
    Global properties of asymptotically de Sitter and Anti de Sitter spacetimes Didier A. Solis May, 17, 2006 arXiv:1803.01171v1 [gr-qc] 3 Mar 2018 Ad Majorem Dei Gloriam. To all those who seek to fulfill God's will in their lives. iii ACKNOWLEDGEMENTS I would like to express my deepest gratitude to: • My advisor Dr. Gregory Galloway, for his constant support and help. This dissertation would not had seen the light of day without his continued encour- agement and guidance. • Dr. N. Saveliev, Dr. M. Cai and Dr. O. Alvarez for their valuable comments and thorough revision of this work. • The Faculty and staff at the Math Department for all the knowledge and affec- tion they shared with me. • CONACYT (Consejo Nacional de Ciencia y Tecnolog´ıa)for the financial support granted to the author. • My family: Douglas, Rosalinda, Douglas Jr, Rosil´uand Josu´efor all their prayers and unconditional love. They are indeed the greatest blessing in my life. • All my friends, especially Guille, for always being there for me to share the struggles and joys of life. • To the CSA gang, for being a living example of faith lived in love and hope. • Last but by no means least, to God, without whom there is no math, laughter or music. iv Contents 1 Introduction 1 1.1 Lorentz vector spaces . 1 1.2 Spacetimes . 6 1.2.1 The Levi-Civita connection . 7 1.2.2 Covariant derivative . 11 1.2.3 Geodesics . 12 1.3 Causal theory . 15 1.3.1 Definitions . 15 1.3.2 Global causality conditions .
    [Show full text]
  • Lecture 8: the Sectional and Ricci Curvatures
    LECTURE 8: THE SECTIONAL AND RICCI CURVATURES 1. The Sectional Curvature We start with some simple linear algebra. As usual we denote by ⊗2(^2V ∗) the set of 4-tensors that is anti-symmetric with respect to the first two entries and with respect to the last two entries. Lemma 1.1. Suppose T 2 ⊗2(^2V ∗), X; Y 2 V . Let X0 = aX +bY; Y 0 = cX +dY , then T (X0;Y 0;X0;Y 0) = (ad − bc)2T (X; Y; X; Y ): Proof. This follows from a very simple computation: T (X0;Y 0;X0;Y 0) = T (aX + bY; cX + dY; aX + bY; cX + dY ) = (ad − bc)T (X; Y; aX + bY; cX + dY ) = (ad − bc)2T (X; Y; X; Y ): 1 Now suppose (M; g) is a Riemannian manifold. Recall that 2 g ^ g is a curvature- like tensor, such that 1 g ^ g(X ;Y ;X ;Y ) = hX ;X ihY ;Y i − hX ;Y i2: 2 p p p p p p p p p p 1 Applying the previous lemma to Rm and 2 g ^ g, we immediately get Proposition 1.2. The quantity Rm(Xp;Yp;Xp;Yp) K(Xp;Yp) := 2 hXp;XpihYp;Ypi − hXp;Ypi depends only on the two dimensional plane Πp = span(Xp;Yp) ⊂ TpM, i.e. it is independent of the choices of basis fXp;Ypg of Πp. Definition 1.3. We will call K(Πp) = K(Xp;Yp) the sectional curvature of (M; g) at p with respect to the plane Πp. Remark. The sectional curvature K is NOT a function on M (for dim M > 2), but a function on the Grassmann bundle Gm;2(M) of M.
    [Show full text]
  • Curves of Constant Curvature and Torsion in the the 3-Sphere
    CURVES OF CONSTANT CURVATURE AND TORSION IN THE 3-SPHERE DEBRAJ CHAKRABARTI, RAHUL SAHAY, AND JARED WILLIAMS Abstract. We describe the curves of constant (geodesic) curvature and torsion in the three-dimensional round sphere. These curves are the trajectory of a point whose motion is the superposition of two circular motions in orthogonal planes. The global behavior may be periodic or the curve may be dense in a Clifford torus embedded in the three- sphere. This behavior is very different from that of helices in three-dimensional Euclidean space, which also have constant curvature and torsion. 1. Introduction Let pM; x; yq be a three dimensional Riemannian manifold, let I Ď R be an open interval, and let γ : I Ñ M be a smooth curve in M, which we assume to be parametrized by the arc length t. It is well-known that the local geometry of γ is characterized by the curvature κ and the torsion τ. These are functions defined along γ and are the coefficients of the well-known Frenet-Serret formulas [2, Vol. IV, pp. 21-23]: D Tptq “ κptqNptq dt D Nptq “ ´κptqTptq `τptqBptq (1.1) dt D Bptq “ ´ τptqNptq; dt where the orthogonal unit vector fields T; N; B, with T “ γ1, along the unit-speed curve D γ, constitute its Frenet frame and dt denotes covariant differentiation along γ with respect to the arc length t. We will assume that each of the functions κ and τ is either nowhere zero or vanishes identically. Additionally, if κ is identically zero, then τ is also taken to be identically zero.
    [Show full text]
  • A Review on Metric Symmetries Used in Geometry and Physics K
    A Review on Metric Symmetries used in Geometry and Physics K. L. Duggala aUniversity of Windsor, Windsor, Ontario N9B3P4, Canada, E-mail address: [email protected] This is a review paper of the essential research on metric (Killing, homothetic and conformal) symmetries of Riemannian, semi-Riemannian and lightlike manifolds. We focus on the main characterization theorems and exhibit the state of art as it now stands. A sketch of the proofs of the most important results is presented together with sufficient references for related results. 1. Introduction The measurement of distances in a Euclidean space R3 is represented by the distance element ds2 = dx2 + dy2 + dz2 with respect to a rectangular coordinate system (x, y, z). Back in 1854, Riemann generalized this idea for n-dimensional spaces and he defined element of length by means of a quadratic 2 i j differential form ds = gijdx dx on a differentiable manifold M, where the coefficients gij are functions of the coordinates system (x1, . , xn), which represent a symmetric tensor field g of type (0, 2). Since then much of the subsequent differential geometry was developed on a real smooth manifold (M, g), called a Riemannian manifold, where g is a positive definite metric tensor field. Marcel Berger’s book [1] includes the major developments of Riemannian geome- try since 1950, citing the works of differential geometers of that time. On the other hand, we refer standard book of O’Neill [2] on the study of semi-Riemannian geometry where the metric g is indefinite and, in particular, Beem-Ehrlich [3] on the global Lorentzian geometry used in relativity.
    [Show full text]
  • Differential Geometry Lecture 18: Curvature
    Differential geometry Lecture 18: Curvature David Lindemann University of Hamburg Department of Mathematics Analysis and Differential Geometry & RTG 1670 14. July 2020 David Lindemann DG lecture 18 14. July 2020 1 / 31 1 Riemann curvature tensor 2 Sectional curvature 3 Ricci curvature 4 Scalar curvature David Lindemann DG lecture 18 14. July 2020 2 / 31 Recap of lecture 17: defined geodesics in pseudo-Riemannian manifolds as curves with parallel velocity viewed geodesics as projections of integral curves of a vector field G 2 X(TM) with local flow called geodesic flow obtained uniqueness and existence properties of geodesics constructed the exponential map exp : V ! M, V neighbourhood of the zero-section in TM ! M showed that geodesics with compact domain are precisely the critical points of the energy functional used the exponential map to construct Riemannian normal coordinates, studied local forms of the metric and the Christoffel symbols in such coordinates discussed the Hopf-Rinow Theorem erratum: codomain of (x; v) as local integral curve of G is d'(TU), not TM David Lindemann DG lecture 18 14. July 2020 3 / 31 Riemann curvature tensor Intuitively, a meaningful definition of the term \curvature" for 3 a smooth surface in R , written locally as a graph of a smooth 2 function f : U ⊂ R ! R, should involve the second partial derivatives of f at each point. How can we find a coordinate- 3 free definition of curvature not just for surfaces in R , which are automatically Riemannian manifolds by restricting h·; ·i, but for all pseudo-Riemannian manifolds? Definition Let (M; g) be a pseudo-Riemannian manifold with Levi-Civita connection r.
    [Show full text]
  • M-Eigenvalues of Riemann Curvature Tensor of Conformally Flat Manifolds
    M-eigenvalues of Riemann Curvature Tensor of Conformally Flat Manifolds Yun Miao∗ Liqun Qi† Yimin Wei‡ August 7, 2018 Abstract We generalized Xiang, Qi and Wei’s results on the M-eigenvalues of Riemann curvature tensor to higher dimensional conformal flat manifolds. The expression of M-eigenvalues and M-eigenvectors are found in our paper. As a special case, M-eigenvalues of conformal flat Einstein manifold have also been discussed, and the conformal the invariance of M- eigentriple has been found. We also discussed the relationship between M-eigenvalue and sectional curvature of a Riemannian manifold. We proved that the M-eigenvalue can determine the Riemann curvature tensor uniquely and generalize the real M-eigenvalue to complex cases. In the last part of our paper, we give an example to compute the M-eigentriple of de Sitter spacetime which is well-known in general relativity. Keywords. M-eigenvalue, Riemann Curvature tensor, Ricci tensor, Conformal invariant, Canonical form. AMS subject classifications. 15A48, 15A69, 65F10, 65H10, 65N22. 1 Introduction The eigenvalue problem is a very important topic in tensor computation theories. Xiang, Qi and Wei [1, 2] considered the M-eigenvalue problem of elasticity tensor and discussed the relation between strong ellipticity condition of a tensor and it’s M-eigenvalues, and then extended the M-eigenvalue problem to Riemann curvature tensor. Qi, Dai and Han [12] introduced the M-eigenvalues for the elasticity tensor. The M-eigenvalues θ of the fourth-order tensor Eijkl are defined as follows. arXiv:1808.01882v1 [math.DG] 28 Jul 2018 j k l Eijkly x y = θxi, ∗E-mail: [email protected].
    [Show full text]
  • LECTURE 9: MORE on CURVATURES 1. Various Types Of
    LECTURE 9: MORE ON CURVATURES 1. Various types of special Riemannian manifolds We shall briefly describe various classes of Riemannian manifolds which are special in terms of curvatures. 1.1. Flat manifolds. Definition 1.1. We say a Riemannian manifold (M; g) is locally Euclidean if near each point p 2 M, there is a coordinate system (U; x1; ··· ; xm) so that on U, g = dx1 ⊗ dx1 + ··· dxm ⊗ dxm: For example, the standard torus Tm, either viewed as the product S1 × · · · × S1 (endow each S1 with the standard rotation-invariant metric dθ ⊗ dθ), or viewed as the quotient Rm=Zm of the standard Euclidean space Rm, is locally Euclidean with an obvious choice of local coordinates. Obviously if (M; g) is locally Euclidean, then the Christoffel symbols associated to the coordinate system alluded above are all zero, thus the curvature tensor of (M; g) vanishes. As a result, (M; g) must be a flat manifold. Conversely, we have Theorem 1.2. If (M; g) is flat if and only if it is a locally Euclidean space. Sketch of proof. One need to prove that any flat manifold (M; g) is locally Euclidean. By using the geometric method that we will develop next time, one can prove that if (M; g) is flat, then one can find, in some neighborhood U of each point p a flat frame, i.e. a set of vector fields X1; ··· ;Xm on U such that •f Xi(q) j 1 ≤ i ≤ mg form a basis of TqM for every q 2 U, •r Y Xi = 0 for all i and for all vector field Y .
    [Show full text]
  • 612 CLASS LECTURE: HYPERBOLIC GEOMETRY 1. Conformal Metrics As a Vector Space, C Has a Canonical Norm, the Same As the Standard
    612 CLASS LECTURE: HYPERBOLIC GEOMETRY JOSHUA P. BOWMAN 1. Conformal metrics As a vector space, C has a canonical norm, the same as the standard R2 norm. Denote this jdzj|one should think of dz as the identity map on C: ζ 7! ζ. This also means C has a standard norm as the tangent space to itself at any of its points. We will consistently use Roman letters (e.g., z) to represent a point in C when we're thinking of it as a topological ◦ space, and Greek letters (e.g., ζ) to represent a point in C thought of as a vector. Let U ⊂ C be a plane domain (non-empty, connected, but not necessarily simply connected). Definition 1.1. A conformal metric on U is a metric of the form ρ jdzj, where ρ is a smooth, positive function on U. We will also use ρ to denote the metric itself, not just the function. Almost always, \conformal" is used with respect to something else; here, we mean that angles in a conformal metric are the same as those in the background Euclidean metric. In other words, the unit ball of the norm in the tangent space at z is invariant under (Euclidean) rotations. Example 1.2. On U = C n f0g, ρ = jdzj=jzj is a conformal metric. U is preserved by multiplication by a non-zero complex number a; call such a map ma. The derivative of ma is again multiplication by a. Let z 2 U and let ζ 2 Tz(U) = C.
    [Show full text]
  • Constant Mean Curvature Spheres in Homogeneous Three-Manifolds
    Constant mean curvature spheres in homogeneous three-manifolds William H. Meeks III∗ Pablo Miray Joaqu´ın Perez´ Antonio Rosz June 29, 2017 Abstract We prove that two spheres of the same constant mean curvature in an arbitrary homoge- neous three-manifold only differ by an ambient isometry, and we determine the values of the mean curvature for which such spheres exist. This gives a complete classification of immersed constant mean curvature spheres in three-dimensional homogeneous manifolds. Mathematics Subject Classification: Primary 53A10, Secondary 49Q05, 53C42 Key words and phrases: Constant mean curvature, Hopf uniqueness, homogeneous three- manifold, Cheeger constant. 1 Introduction In this manuscript we solve the fundamental problem of classifying constant mean curvature spheres in an arbitrary homogeneous three-manifold, where by a sphere, we mean a closed im- mersed surface of genus zero: Theorem 1.1. Let M be a Riemannian homogeneous three-manifold, X denote its Riemannian universal cover, and let Ch(X) denote the Cheeger constant of X. Then, any two spheres in M of the same constant mean curvature differ by an isometry of M. Moreover: (1) If X is not diffeomorphic to R3, then, for every H R, there exists a sphere of constant mean curvature H in M. 2 arXiv:1706.09394v1 [math.DG] 28 Jun 2017 (2) If X is diffeomorphic to R3, then the values H R for which there exists a sphere of constant mean curvature H in M are exactly those with2H > Ch(X)=2. j j ∗This material is based upon work for the NSF under Award No.
    [Show full text]