Information Geometry of Bivariate Gamma Exponential Distributions
|
|
- Gary Reynolds
- 6 years ago
- Views:
Transcription
1 Information Geometry of Bivariate Gamma Exponential Distriutions Khadiga Ali Arwini Department of Mathematics Al-Fateh University Tripoli-Liya July 7, 9 Astract This paper is devoted to the information geometry of the family of ivariate gamma exponential distriutions, which have gamma and Pareto margins, and discuss some of its applications. We egin y considering the parameter ivariate gamma exponential manifold as a Riemannian -manifold; y following the Roa s idea to use the Fisher information matrix (FIM), and derive the α-geometry as; α-connections, α-curvature tensor, α-ricci curvature with its eigenvalues and eigenvectors, and α-scalar curvature; where the -geometry is corresponds to the geometry induced y the Levi-Civita connection, and we show that this space has a non-constant negative scalar curvature. In addition, we consider four sumanifolds as special cases, and discuss their geometrical structures, and we prove that one of these sumanifolds is an isometric isomorph of the univariate gamma manifold. Then we introduce log-ivariate gamma exponential distriutions, which have log-gamma and log-pareto margins, and we show that this family of distriutions determines a Riemannian -manifold which is isometry with the origin manifold. AMS Suject Classification (): 5B, 6G5 Key words: information geometry, statistical manifold, gamma distriution, Pareto distriution, ivariate distriutions, ivariate gamma exponential distriution Introduction Statistical manifolds are representations of families of proaility distriutions that allow differential geometric methods to e applied to prolems in stochastic processes, mathematical statistics and information theory. The origin work was due to Rao [], who considered a family of proaility distriutions as Riemannian manifold using the Fisher information matrix (FIM) as a Riemannian metric. In 975, Efron [6] defined the curvature in statistical manifolds, and gave a statistical interpretation for the curvature with application to second order efficiency. Then Amari [] introduced a one-parameter family of affine connections (α-connection), which contain the Levi-Civita connection of the Fisher metric when α =, and turn out to have importance and are part of larger systems of connections for which staility results follow [5]. He further proposed a differential-geometrical framework for constructing a higherorder asymptotic theory of statistical inference. He, defining the α-curvature of a sumanifold, pointed out important roles of the exponential and mixture curvatures and their duality in statistical inference. Several researchers studied the information geometry and its applications for some families of univariate and multivariate distriutions. Amari [] showed that the family of univariate Gaussian distriutions has a constant negative curvature. Oller [?] studied the geometry of the extreme value distriutions as, Gumel, Weiull and Frethet distriutions, and he showed that all these spaces are with constant negative curvatures, and he otained the geodesic distances in each case. Gamma manifold studied y many researcher eg [], and Dodson [] proved that every neighourhood of an exponential distriution contains a neighourhood of gamma distriutions, using an information theoretic metric topology. Adel- All, Mahmoud and Add-Ellah [] showed that the Pareto family is a space with constant positive curvature and they otained the geodesics, and they showed the relation etween the geodesic distance and the J-divergence.
2 Information Geometry of Bivariate Gamma Exponential Distriutions For ivariate and multivariate distriutions, the Fisher information metrics or the geometrical structures for most of these ivariate and multivariate distriutions has een found to e intractale; the only exceptions have een the families as, the ivariate and multivariate Gaussian distriutions, the Mckay ivariate gamma distriutions, and the Freund ivariate exponential distriutions. For a summary of ivariate and multivariate distriutions see Kotz, Balakrishnan and Johnson [7]. Sato, Sugawa and Kawaguchi [] proved that the 5-dimensional space of ivariate Gaussian distriutions has a negative constant curvature and if the covariance is zero, the space ecomes an Einstein space. After that, Skovgaard [] in 984, extended this work to the family of multivariate Gaussian model, and provided the geodesic distances. The geometry for Mckay manifold where the variales are positively correlated, otained y Arwini and Dodson [], who showed that this space can provide a metrization of departures from randomness and independence for ivariate processes, also they used the geometry of Mckay manifold in some application in hydrology, concerns tomographic images of soil samples in hydrology surveys, see [4]. After that, the geometry for Freund 4-manifold investigated y Arwini and Dodosn [], and they showed the important of this family ecause exponential distriutions represent intervals etween events for Poisson processes and Freund distriutions can model ivariate processes with positive and negative covariance. In this paper we are interested to study the Becker and Roux s and Steel and le Roux s ivariate gamma distriution, with gamma margins, which are ased on plausile physical models, and they also include the Freund distriution as a special case, see [7], the Fisher information matrix is found to e intractale. However, Steel and Roux in 989 [] studied compound distriutions of this ivariate gamma, and they otain the ivariate gamma exponential distriution with gamma and Pareto margins, hence we consider this family of distriution as a Riemannian -manifold and discuss its geometry. Bivariate gamma exponential distriutions The ivariate gamma exponential distriution has event space Ω = R + R + and proaility density function (pdf) given y f(x, y; a,, c) = a c Γ() x e (a x+c x y) for a,, c R + (.) where Γ is the gamma function which has the formula Γ(z) = t z e t dt. Figure show surface and control plots of ivariate gamma exponential distriution f(x, y; a,, c) for the range x [, 5] and y [, ], with unit parameters a = = c =.. Margins and correlation The marginal distriutions, of X and Y are gamma distriution and Pareto distriution, respectively. f X (x) = f Y (y) = a Γ() x e a x, x > (.) c a, y >. (.) + (a + c y) While the conditional marginal function f Y X given x is exponential distriution, with density function: f Y X (y x) = (c x) e (c x) y, y >. (.4)
3 Khadiga Ali Arwini y x Figure : The ivariate gamma exponential family of densities with parameters a = = c = as a surface and as a contour plot, shown for the range x [, 5] and y [, ]. The covariance and correlation coefficient of X and Y are given y: Cov(X, Y ) = c ( ), > Γ() ρ(x, Y ) =, > + Γ( + ) Γ( + ) Note that the ivariate gamma exponential distriution does not contains the independent case, with negative correlation which depends on only of the parameter. The correlation ρ decreases from to.5 where < < 4, and increases from.5 to where > 4. See figure.. Log-likelihood function and Shannon s entropy The log-likelihood function for the ivariate gamma exponential distriution (.) is l(x, y; a,, c) = log(f(x, y; a,, c)) = x (a + c y) + log(c) + (log(a) + log(x)) log(γ()). By direct calculation Shannon s information theoretic entropy for the ivariate gamma exponential distriution, which is the negative of the expectation of the log-likelihood function, is given y S f (a,, c) = l(x, y; a,, c) f(x, y; a,, c) dx dy, = + log(c) + log(γ()) ψ() (.5) Note that S f (a,, c) is independent of the parameter a. In the case when = which is the random case for the marginal function f X, the Shannon s entropy is S f (a,, c) = log(c) + γ where γ is the Euler gamma, in this case S f tends to e zero when c = e +γ. When c =, the Shannon s entropy is
4 4 Information Geometry of Bivariate Gamma Exponential Distriutions ρ(x, Y ) Figure : The correlation coefficient ρ(x, Y ) etween the variales X and Y as a function of the parameter, for the range (, 7). Note that the minimum value of ρ is.5 and occurs at = 4. S f (a,, c) = + log(γ()) ψ(). In this case S f tends to e zero when 6.5. Figure shows a plot of S f (a,, c) in the domain, c (, ), and figure 4 show plots of S f (a,, c) for the cases when = and c =.. Natural coordinate system and potential function We can rewrite the ivariate gamma exponential density funtion (.) as f(x, y; a,, c) = = a c Γ() x e (a x+c x y), = e a ( x)+c ( x y)+ (log(x)) (log(γ()) log(a) log(c)) (.6) Hence the family of ivariate gamma exponential distriutions forms an exponential family, with natural coordinate system (θ) = (θ, θ, θ ) = (a,, c), random variale (x, x, x ) = (F (x), F (x), F (x)) = ( x, x y, log x), and potential function ϕ(a,, c) = log(γ()) log(a) log(c)..4 Fisher information matrix FIM The Fisher Information (FIM) is given y the expectation of the covariance of partial derivatives of the log-likelihood function. Here the Fisher information metric components of the family of ivariate gamma exponential distriutions M with coordinate system (θ) = (θ, θ, θ ) = (a,, c) are given y g ij = = ϕ θ i θ j l(x, y, θ) θ i θ j f(x, y) dy dx, Hence, g = [g ij ] = a a a ψ () c. (.7)
5 Khadiga Ali Arwini c Figure : A surface and a contour plot for the Shannon s information entropy S f, for ivariate gamma exponential distriutions in the domain, c (, ). where ψ() = Γ () Γ() is the digamma function. and the variance covariance matrix is g = [g ij ] = a ψ () ψ () a ψ () a ψ () ψ () c. (.8) Geometry of the ivariate gamma exponential manifold. Bivariate gamma exponential manifold Let M e the family of all ivariate gamma exponential distriutions M = {f(x, y; a,, c) = a c Γ() x e (a x+c x y) a,, c R + }, x, y R + (.9) so the parameter space is R + R + R + and the random variales are (x, y) Ω = R + R +. We can consider M as a Riemannian -manifold with coordinate system (θ, θ, θ ) = (a,, c) and Fisher information metric g (.7), which is ds g = a da + ψ () d + c dc + da d. a. α-connections For each α R, the α (or (α) )-connection is the torsion-free affine connection with components ( log f log f ij,k = + α ) log f log f log f f dy dx θ i θ j θ k θ i θ j θ k
6 6 Information Geometry of Bivariate Gamma Exponential Distriutions c Figure 4: On the left, Shannon s information entropy S f for ivariate gamma exponential distriutions with unit parameter β =, in the domain c (, ). Note that S f tends to e zero when c = e +γ. On the right, Shannon s information entropy S f for ivariate gamma exponential distriutions with unit parameter c =, in the domain (, 5). Note that S f tends to e zero when 6.5. We have an affine connection (α) defined y: So y solving the equations we otain the components of (α). (α) i j, k = ij,k, where i = θ i ij,k = h= g kh Γ h(α) ij, (k =,, ). Here we give the analytic expressions for the α-connections with respect to coordinates (θ, θ, θ ) = (a,, c) : The nonzero independent components ij,k are and the components Γ i(α) jk, = (α ) a,, =, = (α ) Γ(α), = a,, = (α ) ψ (),, = α c. (.) of the (α) -connections are given y = [ ij ] = (α ) ( ψ () ) a ( ψ () ) (α ) ψ () ψ () (α ) ψ () ψ () a (α ) ψ () ψ (). (.)
7 Khadiga Ali Arwini 7 = [ ij ] = (α ) a ( ψ () ) α a ψ () a α a ψ () a (α ) ψ () ψ (). (.) = [ ij ] = α c. (.). α-curvatures By direct calculation we provide various α-curvature ojects of the ivariate gamma exponential manifold M, as: the α-curvature tensor, the α-ricci curvature, the α-scalar curvature, the α-sectional curvature, and the α-mean curvature. The α-curvature tensor components, which are defined as: ( R (α) ijkl = g hl i Γ h(α) jk j Γ h(α) ik + h= m= Γ h(α) im Γm(α) jk Γ h(α) jm Γm(α) ik ), (i, j, k, l =,, ) are given y while the other independent components are zero. ( α R (α) = R (α) = ) (ψ () + ψ ()) 4 a ( ψ. (.4) () ) Contracting R (α) ijkl with gil we otain the components R (α) jk of the α-ricci tensor R (α) = [R (α) jk ] = (α ) (ψ ()+ ψ ()) (α ) (ψ ()+ ψ ()) 4 a ( ψ () ) 4 a ( ψ () ) 4 a ( ψ () ) 4 ( ψ () ) (α ) (ψ ()+ ψ ()) (α ) ψ () (ψ ()+ ψ ()). (.5) The eigenvalues and the eigenvectors of the α-ricci tensor are given in the case when α =, y (+a ψ ()) ( ψ () ) (ψ ()+ ψ ()) ( ψ () ) (4 a + a ψ ()+a 4 ψ () ) (ψ ()+ ψ ()) 8 a ( ψ () ) (+a ψ ()) ( ψ () ) (ψ ()+ ψ ())+ ( ψ () ) (4 a + a ψ ()+a 4 ψ () ) (ψ ()+ ψ ()) 8 a ( ψ () ) (.6) a + a ψ () a + a ψ () + ( ψ () ) (4 a + a ψ ()+a 4 ψ () ) (ψ ()+ ψ ()) a( ψ () ) (ψ ()+ ψ ()) ( ψ () ) (4 a + a ψ ()+a 4 ψ () ) (ψ ()+ ψ ()) a ( ψ () ) (ψ ()+ ψ ()) (.7) By contracting the Ricci curvature components R (α) ij the scalar curvature R (α) : with the inverse metric components g ij we otain
8 8 Information Geometry of Bivariate Gamma Exponential Distriutions - 4 R (α) R () - α 4 - Figure 5: The scalar curvature R (α) for ivariate gamma exponential manifold, in the range α [, ] and (, 4) on the left, and α = and (, 4) on the right. R () is negative increasing function of the parameter from to as tends to. R (α) = ( α ) (ψ () + ψ ()) ( ψ () ). (.8) Note that the α-scalar curvature R (α) is a function of only the parameter, and in the case when α = the scalar curvature R () is negative and its increases from to. Figure?? shows the scalar curvature R (α) in the range α [, ] and (, 4), and the -scalar curvature R () in the range (, 4). The α-mean curvatures ϱ (α) (λ) (λ =,, ) are ϱ (α) () = ϱ (α) () = ( α ) (ψ () + ψ ()) ( ψ () ), ϱ (α) () =. (.9) The non-zero component of the α-sectional curvatures ϱ (α) (λ, µ) (λ, µ =,, ) is ϱ (α) (, ) = ( α ) (ψ () + ψ ()) 4 ( ψ () ). (.).4 Sumanifolds.4. Sumanifold M M: = In the case when =, the ivariate gamma exponential distriution (.) reduces to the form f(x, y; a, c) = a c x e (a x+c x y), x >, y >, a, c >. (.)
9 Khadiga Ali Arwini 9 when the x and y margins have the exponential distriution and Pareto distriution (with unit shape parameter), respectively: f X (x) = a e a x, x > (.) c a f Y (y) =, y >. (.) (a + c y) The Fisher metric with respect to the coordinate system (θ, θ ) = (a, c) is g = [ a c ]. (.4) Here we note that, the sumanifold M and the family of all independent ivariate exponential distriutions, which are the direct product of two exponential distriutions a e a x c e c y have the same Fisher metric. Hence the sumanifold M is an isometry of the manifold of independent ivariate exponential distriutions. In this manifold all the curvatures are zero, while the nonzero independent components ij,k and Γi(α) jk of the (α) -connections are.4. Sumanifold M M: c =, = α a,, = α c, = α, a = α. (.5) c In the case when c =, the ivariate gamma exponential distriution (.) reduces to the form f(x, y; a, ) = a Γ() x e (a x+x y), x >, y >, a, >. (.6) The x and y margins have the exponential and Pareto distriutions, respectively: f X (x) = f Y (y) = a Γ() x e a x, x > (.7) a, y >. (.8) + (a + y) where M and M have the same correlation coefficient (.). The Fisher metric with respect to the coordinate system (θ, θ ) = (a, ) is g = [ a a a ψ () ]. (.9)
10 Information Geometry of Bivariate Gamma Exponential Distriutions Note that the sumanifold M and the gamma manifold, which is the family of all univariate gamma proaility densities functions a Γ() x e a x, x >, a, > have the same Fisher metric. Hence, the family of ivariate gamma exponential proaility density functions with unit parameter c =, is an isometry of the gamma manifold with information theoretical metric topology. Here we give the geometrical structures of the sumanifold M. The nonzero independent components of the (α) -connections are given y:, = (α ) a,, =, = (α ) Γ(α), = a,, = (α ) ψ () (.) = (α ) ( ψ () ) a ( ψ () ) (α ) ψ () ψ () (α ) ψ () ψ () a (α ) ψ () ψ (). (.) = [ (α ) ( α) a ( ψ () ) a ψ () a ( α) a ψ () a (α ) ψ () ψ () ]. (.) The α-curvature tensor is given y while the other independent components are zero. The α-ricci curvature is R (α) = ( α R (α) = R (α) = ) (ψ () + ψ ()) 4 a ( ψ. (.) () ) (α ) (ψ ()+ ψ ()) (α ) (ψ ()+ ψ ()) 4 a ( ψ () ) 4 a ( ψ () ) (α ) (ψ ()+ ψ ()) (α ) ψ () (ψ ()+ ψ ()) 4 a ( ψ () ) 4 ( ψ () ). (.4) The α-scalar curvature is R (α) = ( α ) (ψ () + ψ ()) ( ψ () ). (.5) Note that M and M have the same scalar curvature, which is a negative non constant in the case when α = ; and this has limiting value as and as..4. Sumanifold M M: = In the case when =, the ivariate gamma exponential distriution (.) reduces to the form f(x, y; a, c) = a c x e (a x+c x y), x >, y >, a, c >. (.6) The x and y margins have the gamma distriution(with shape parameter= ) and Pareto distriution (with shape parameter=), respectively: f X (x) = a x e a x, x > (.7) f Y (y) = c a, y >. (.8) (a + c y)
11 Khadiga Ali Arwini The Fisher metric with respect to the coordinate system (θ, θ ) = (a, c) is g = [ a c The nonzero independent components of the (α) -connections are given y: ]. (.9) While all the curvatures are zero., =, = (α ) a, (α ) c, = α, a = α. (.4) c.4.4 Sumanifold M 4 M: a = c In the case when a = c, the ivariate gamma exponential distriution (.) reduces to the form f(x, y; a, ) = a+ Γ() x e a(x+x y), x >, y >, a, >. (.4) The x and y margins have the gamma and Pareto distriutions, respectively: f X (x) = f Y (y) = a Γ() x e a x, x > (.4), y >. (.4) + ( + y) and the correlation coefficient is given y ρ(x, Y ) = Γ() ( ) ( ) (Γ() Γ(+)) Γ(+) (Γ(+) Γ(+)), >. (.44) With respect to the coordinate system (θ, θ ) = (a, ) the Fisher metric is [ + ] g = a a ψ. (.45) () The nonzero independent components of the (α) -connections are given y: a, = (α ) ( + ) a,, =, = (α ) Γ(α), = a,, = (α ) ψ (). (.46) = (α ) ( (+) ψ () ) a ((+) ψ () ) (α ) ψ () (+) ψ () (α ) ψ () (+) ψ () a (α ) ψ () (+) ψ (). (.47)
12 Information Geometry of Bivariate Gamma Exponential Distriutions m n Figure 6: The log-ivariate gamma exponential family of densities with parameters a = = c = as a surface and as a contour plot, shown for the range n, m [, ]. = [ (+) (α ) a ((+) ψ () ) (α ) a ((+) ψ () ) (α ) a ((+) ψ () ) (+) (α ) ψ () (+) ψ () ]. (.48) The α-curvature tensor is given y ( α R (α) = R (α) = ) (ψ () + ( + ) ψ ()) 4 a (( + ) ψ. (.49) () ) while the other independent components are zero. α-ricci curvature is R (α) = (+) (α ) (ψ ()+(+) ψ ()) (α ) (ψ ()+(+) ψ ()) 4 a ((+) ψ () ) 4 a ((+) ψ () ) (α ) (ψ ()+(+) ψ ()) (α ) ψ () (ψ ()+(+) ψ ()) 4 a ((+) ψ () ) 4 ((+) ψ () ). (.5) The scalar curvature R () is given y: where the scalar curvature R () is negative non constant. R (α) = ( α ) (ψ () + ( + ) ψ ()) (( + ) ψ () ). (.5).5 Log-ivariate gamma exponential manifold Here we introduce a log-ivariate gamma exponential distriution, which arises from the ivariate gamma exponential distriution (.) for non-negative random variales n = e x and m = e y. So log-ivariate gamma exponential distriution, has proaility density function g(n, m) = a c n a ( log(n)) m Γ() e c log(n) log(m), < n <, < m <. (.5)
13 Khadiga Ali Arwini where a,, c >. Figure 6 show plots of the log-ivariate gamma exponential family of densities with unit parameters a, and c. Note that the margins of N and M are log-gamma distriution and log-pareto distriution, respectively. f N (n) = a na ( a log(n)) Γ() f M (m) =, < n < ( ) a c m (a c log(m)) +, < m < The covariance and correlation coefficient of the variales N and M are given y: Cov(N, M) = e a c ( a ρ(n, M) = ( ( + a) ( ) ( a e + a c E( +, ) ( ) E( +, a ) + a c c ), a +a ) a (+a) ) ( e c E( +, +a ) c ) ( ) E( +, a c ) ( ) (. c c Γ(, a c ) ( ) a Γ(, a c )) where E(n, z) is the exponential integral function, which is defined y e z t t dt. n This family of densities determines a Riemannian -manifold which is isometry with the ivariate gamma exponential manifold M. 4 Applications Still????? Acknowledgement The author would like to thank Professor Saralees Nadarajah for helpful comments concerning ivariate gamma distriutions. References [] N. H. Adel-All, M. A. W. Mahmoud and H. N. Ad-Ellah. Geomerical Properties of Pareto Distriution. Applied Mathematics and Computation 45() -9. [] Amari S. and Nagaoka H., Methods of Information Geometry, American Mathematical Society, Oxford University Press, [] Khadiga Arwini and C.T.J. Dodson. Information Geometry Near Randomness and Near Independence. Lecture Notes in Mathematics 95, Springer-Verlag, Berlin, Heidelerg, New York 8. [4] Khadiga Arwini, C.T.J. Dodson, S. Felipussi and J. Scharcanski. Information Geometric Classification of Porous Media Tomographic Images. Preprint (). [5] C.T.J. Dodson. Systems of connections for parametric models. In Proc. Workshop on Geometrization of Statistical Theory 8- Octoer 987. Ed. C.T.J. Dodson, ULDM Pulications, University of Lancaster, 987, pp 5-7.
14 4 Information Geometry of Bivariate Gamma Exponential Distriutions [6] B. Efron. Defining the curvature of a statistical prolem (with application to second order efficiency) (with discussion). Ann. Statist.,, (975) [7] Samuel Kotz, N. Balakrishnan and Norman L. Johnson. Continuous Multivariate distriutions. Volume : Models and Applications, Second Edition, John Wiley and Sons, New York,. [8] Saralees Nadarajah. The Bivariate Gamma Exponential Distriution With Application To Drought Data. Summited to Korean Society for Computational and Applied Mathematics and Korean SIGCAM. [9] Linyu Peng, Huafei Sun and Lin Jiu. The Geometric Structure of the Pareto Distriution. Boletiin de la Asociacion Matematica Venezolana, Vol. XIV, Nos. y (7) 5. [] C.R. Rao. Information and accuracy attainale in the estimation of statistical parameters. Bull. Calcutta Math. Soc. 7, (945) 8-9. [] Y.Sato, K.Sugawa and M.Kawaguchi. The geometrical structure of the parameter space of the two-dimensional normal distriution. Division of information engineering, Hokkaido University, Sapporo, Japan (977). [] L.T. Skovgaard. A Riemannian geometry of the multivariate normal model. Scand. J. Statist. (984) -. [] Steel, S. J., and le Roux, N. J.(989). A class of compound distriutions of the reparametrised ivariate gamma extension. South African Statistical Jouranl,, -4. [4] S. Yue. A Bivariate Gamma Distriution for Use in Multivariate Flood Frequency Analysis. Hydrological Processes 5(), -45. [5] S. Yue, T. B. M. J. Ouarda and B. Boee. A Review of Bivariate Gamma Distriutionks for Hydrological Applicaiton. Journal of Hydrlogy 46(), -8.
Neighbourhoods of Randomness and Independence
Neighbourhoods of Randomness and Independence C.T.J. Dodson School of Mathematics, Manchester University Augment information geometric measures in spaces of distributions, via explicit geometric representations
More informationInformation geometry for bivariate distribution control
Information geometry for bivariate distribution control C.T.J.Dodson + Hong Wang Mathematics + Control Systems Centre, University of Manchester Institute of Science and Technology Optimal control of stochastic
More informationOn the entropy flows to disorder
CHAOS 2009 Charnia, Crete 1-5 June 2009 On the entropy flows to disorder C.T.J. Dodson School of Mathematics University of Manchester, UK Abstract Gamma distributions, which contain the exponential as
More informationOn the entropy flows to disorder
On the entropy flows to disorder C.T.J. Dodson School of Mathematics, University of Manchester, Manchester M 9PL, UK ctdodson@manchester.ac.uk May, 9 Abstract Gamma distributions, which contain the exponential
More informationarxiv: v2 [math-ph] 22 May 2009
On the entropy flows to disorder arxiv:8.438v [math-ph] May 9 C.T.J. Dodson School of Mathematics, University of Manchester, Manchester M3 9PL, UK ctdodson@manchester.ac.uk August 6, 8 Abstract Gamma distributions,
More informationUniversal connection and curvature for statistical manifold geometry
Universal connection and curvature for statistical manifold geometry Khadiga Arwini, L. Del Riego and C.T.J. Dodson Department of Mathematics University of Manchester Institute of Science and Technology
More informationUniversal connection and curvature for statistical manifold geometry
Universal connection and curvature for statistical manifold geometry Khadiga Arwini, L. Del Riego and C.T.J. Dodson School of Mathematics, University of Manchester, Manchester M60 QD, UK arwini00@yahoo.com
More informationAN AFFINE EMBEDDING OF THE GAMMA MANIFOLD
AN AFFINE EMBEDDING OF THE GAMMA MANIFOLD C.T.J. DODSON AND HIROSHI MATSUZOE Abstract. For the space of gamma distributions with Fisher metric and exponential connections, natural coordinate systems, potential
More informationOn the entropy flows to disorder
On the entropy flows to disorder School of Mathematics University of Manchester Manchester M 9PL, UK (e-mail: ctdodson@manchester.ac.uk) C.T.J. Dodson Abstract. Gamma distributions, which contain the exponential
More informationInformation geometry of the power inverse Gaussian distribution
Information geometry of the power inverse Gaussian distribution Zhenning Zhang, Huafei Sun and Fengwei Zhong Abstract. The power inverse Gaussian distribution is a common distribution in reliability analysis
More informationTopics in Information Geometry. Dodson, CTJ. MIMS EPrint: Manchester Institute for Mathematical Sciences School of Mathematics
Topics in Information Geometry Dodson, CTJ 005 MIMS EPrint: 005.49 Manchester Institute for Mathematical Sciences School of Mathematics The University of Manchester Reports available from: And by contacting:
More informationLecture Notes in Mathematics Editors: J.-M. Morel, Cachan F. Takens, Groningen B. Teissier, Paris
Lecture Notes in Mathematics 1953 Editors: J.-M. Morel, Cachan F. Takens, Groningen B. Teissier, Paris Khadiga A. Arwini Christopher T.J. Dodson Information Geometry Near Randomness and Near Independence
More informationComparing Distance Measures Between Bivariate Gamma Processes
Comparing Distance easures Between Bivariate Gamma Processes Khadiga Arwini 1, C.T.J. Dodson 1, S. Felipussi and J. Scharcanski 1 School of athematics, University of anchester, anchester 6 1QD, UK Departamento
More informationIncluding colour, pdf graphics and hyperlinks
Including colour, pdf graphics and hyperlinks CTJ Dodson School of Mathematics, University of Manchester, Manchester M3 9PL, UK Summer Abstract You are unlikely to want your research articles to use coloured
More informationSymplectic and Kähler Structures on Statistical Manifolds Induced from Divergence Functions
Symplectic and Kähler Structures on Statistical Manifolds Induced from Divergence Functions Jun Zhang 1, and Fubo Li 1 University of Michigan, Ann Arbor, Michigan, USA Sichuan University, Chengdu, China
More informationInformation geometry of Bayesian statistics
Information geometry of Bayesian statistics Hiroshi Matsuzoe Department of Computer Science and Engineering, Graduate School of Engineering, Nagoya Institute of Technology, Nagoya 466-8555, Japan Abstract.
More informationChapter 5 continued. Chapter 5 sections
Chapter 5 sections Discrete univariate distributions: 5.2 Bernoulli and Binomial distributions Just skim 5.3 Hypergeometric distributions 5.4 Poisson distributions Just skim 5.5 Negative Binomial distributions
More informationNOTE ON ASYMPTOTICALLY CONICAL EXPANDING RICCI SOLITONS
PROCEEDINGS OF THE AMERICAN MATHEMATICAL SOCIETY Volume 00, Number 0, Pages 000 000 S 0002-9939(XX)0000-0 NOTE ON ASYMPTOTICALLY CONICAL EXPANDING RICCI SOLITONS JOHN LOTT AND PATRICK WILSON (Communicated
More informationIntroduction to Information Geometry
Introduction to Information Geometry based on the book Methods of Information Geometry written by Shun-Ichi Amari and Hiroshi Nagaoka Yunshu Liu 2012-02-17 Outline 1 Introduction to differential geometry
More informationApplications of Fisher Information
Applications of Fisher Information MarvinH.J. Gruber School of Mathematical Sciences Rochester Institute of Technology 85 Lomb Memorial Drive Rochester,NY 14623 jgsma@rit.edu www.people.rit.edu/mjgsma/syracuse/talk.html
More informationInformation Geometry: Background and Applications in Machine Learning
Geometry and Computer Science Information Geometry: Background and Applications in Machine Learning Giovanni Pistone www.giannidiorestino.it Pescara IT), February 8 10, 2017 Abstract Information Geometry
More informationF -Geometry and Amari s α Geometry on a Statistical Manifold
Entropy 014, 16, 47-487; doi:10.3390/e160547 OPEN ACCESS entropy ISSN 1099-4300 www.mdpi.com/journal/entropy Article F -Geometry and Amari s α Geometry on a Statistical Manifold Harsha K. V. * and Subrahamanian
More informationRiemannian geometry of positive definite matrices: Matrix means and quantum Fisher information
Riemannian geometry of positive definite matrices: Matrix means and quantum Fisher information Dénes Petz Alfréd Rényi Institute of Mathematics Hungarian Academy of Sciences POB 127, H-1364 Budapest, Hungary
More informationCOMPLETE GRADIENT SHRINKING RICCI SOLITONS WITH PINCHED CURVATURE
COMPLETE GRADIENT SHRINKING RICCI SOLITONS WITH PINCHED CURVATURE GIOVANNI CATINO Abstract. We prove that any n dimensional complete gradient shrinking Ricci soliton with pinched Weyl curvature is a finite
More informationElements of differential geometry
Elements of differential geometry R.Beig (Univ. Vienna) ESI-EMS-IAMP School on Mathematical GR, 28.7. - 1.8. 2014 1. tensor algebra 2. manifolds, vector and covector fields 3. actions under diffeos and
More informationSome illustrations of information geometry in biology and physics
Some illustrations of information geometry in biology and physics C.T.J. Dodson School of Mathematics, University of Manchester UK Abstract Many real processes have stochastic features which seem to be
More informationInformation geometry of mirror descent
Information geometry of mirror descent Geometric Science of Information Anthea Monod Department of Statistical Science Duke University Information Initiative at Duke G. Raskutti (UW Madison) and S. Mukherjee
More informationarxiv: v3 [math.dg] 13 Mar 2011
GENERALIZED QUASI EINSTEIN MANIFOLDS WITH HARMONIC WEYL TENSOR GIOVANNI CATINO arxiv:02.5405v3 [math.dg] 3 Mar 20 Abstract. In this paper we introduce the notion of generalized quasi Einstein manifold,
More informationCS Lecture 19. Exponential Families & Expectation Propagation
CS 6347 Lecture 19 Exponential Families & Expectation Propagation Discrete State Spaces We have been focusing on the case of MRFs over discrete state spaces Probability distributions over discrete spaces
More informationDIFFERENTIAL GEOMETRY. LECTURE 12-13,
DIFFERENTIAL GEOMETRY. LECTURE 12-13, 3.07.08 5. Riemannian metrics. Examples. Connections 5.1. Length of a curve. Let γ : [a, b] R n be a parametried curve. Its length can be calculated as the limit of
More informationSUBTANGENT-LIKE STATISTICAL MANIFOLDS. 1. Introduction
SUBTANGENT-LIKE STATISTICAL MANIFOLDS A. M. BLAGA Abstract. Subtangent-like statistical manifolds are introduced and characterization theorems for them are given. The special case when the conjugate connections
More informationTHE BIVARIATE F 3 -BETA DISTRIBUTION. Saralees Nadarajah. 1. Introduction
Commun. Korean Math. Soc. 21 2006, No. 2, pp. 363 374 THE IVARIATE F 3 -ETA DISTRIUTION Saralees Nadarajah Abstract. A new bivariate beta distribution based on the Appell function of the third kind is
More informationCurvature-homogeneous spaces of type (1,3)
Curvature-homogeneous spaces of type (1,3) Oldřich Kowalski (Charles University, Prague), joint work with Alena Vanžurová (Palacky University, Olomouc) Zlatibor, September 3-8, 2012 Curvature homogeneity
More informationFisica Matematica. Stefano Ansoldi. Dipartimento di Matematica e Informatica. Università degli Studi di Udine. Corso di Laurea in Matematica
Fisica Matematica Stefano Ansoldi Dipartimento di Matematica e Informatica Università degli Studi di Udine Corso di Laurea in Matematica Anno Accademico 2003/2004 c 2004 Copyright by Stefano Ansoldi and
More information7 The cigar soliton, the Rosenau solution, and moving frame calculations
7 The cigar soliton, the Rosenau solution, and moving frame calculations When making local calculations of the connection and curvature, one has the choice of either using local coordinates or moving frames.
More informationOn Einstein Nearly Kenmotsu Manifolds
International Journal of Mathematics Research. ISSN 0976-5840 Volume 8, Number 1 (2016), pp. 19-24 International Research Publication House http://www.irphouse.com On Einstein Nearly Kenmotsu Manifolds
More informationRigidity and Non-rigidity Results on the Sphere
Rigidity and Non-rigidity Results on the Sphere Fengbo Hang Xiaodong Wang Department of Mathematics Michigan State University Oct., 00 1 Introduction It is a simple consequence of the maximum principle
More informationThe volume growth of complete gradient shrinking Ricci solitons
arxiv:0904.0798v [math.dg] Apr 009 The volume growth of complete gradient shrinking Ricci solitons Ovidiu Munteanu Abstract We prove that any gradient shrinking Ricci soliton has at most Euclidean volume
More informationRiemannian Curvature Functionals: Lecture I
Riemannian Curvature Functionals: Lecture I Jeff Viaclovsky Park City athematics Institute July 16, 2013 Overview of lectures The goal of these lectures is to gain an understanding of critical points of
More informationHigher dimensional Kerr-Schild spacetimes 1
Higher dimensional Kerr-Schild spacetimes 1 Marcello Ortaggio Institute of Mathematics Academy of Sciences of the Czech Republic Bremen August 2008 1 Joint work with V. Pravda and A. Pravdová, arxiv:0808.2165
More informationDraft version September 15, 2015
Novi Sad J. Math. Vol. XX, No. Y, 0ZZ,??-?? ON NEARLY QUASI-EINSTEIN WARPED PRODUCTS 1 Buddhadev Pal and Arindam Bhattacharyya 3 Abstract. We study nearly quasi-einstein warped product manifolds for arbitrary
More informationAn introduction to General Relativity and the positive mass theorem
An introduction to General Relativity and the positive mass theorem National Center for Theoretical Sciences, Mathematics Division March 2 nd, 2007 Wen-ling Huang Department of Mathematics University of
More informationMATH DIFFERENTIAL GEOMETRY. Contents
MATH 3968 - DIFFERENTIAL GEOMETRY ANDREW TULLOCH Contents 1. Curves in R N 2 2. General Analysis 2 3. Surfaces in R 3 3 3.1. The Gauss Bonnet Theorem 8 4. Abstract Manifolds 9 1 MATH 3968 - DIFFERENTIAL
More informationChapter 2 Exponential Families and Mixture Families of Probability Distributions
Chapter 2 Exponential Families and Mixture Families of Probability Distributions The present chapter studies the geometry of the exponential family of probability distributions. It is not only a typical
More informationCurvature homogeneity of type (1, 3) in pseudo-riemannian manifolds
Curvature homogeneity of type (1, 3) in pseudo-riemannian manifolds Cullen McDonald August, 013 Abstract We construct two new families of pseudo-riemannian manifolds which are curvature homegeneous of
More informationDifferential Geometry II Lecture 1: Introduction and Motivation
Differential Geometry II Lecture 1: Introduction and Motivation Robert Haslhofer 1 Content of this lecture This course is on Riemannian geometry and the geometry of submanifol. The goal of this first lecture
More informationA Characterization of Einstein Manifolds
Int. J. Contemp. Math. Sciences, Vol. 7, 212, no. 32, 1591-1595 A Characterization of Einstein Manifolds Simão Stelmastchuk Universidade Estadual do Paraná União da Vitória, Paraná, Brasil, CEP: 846- simnaos@gmail.com
More informationMATHEMATICAL ENGINEERING TECHNICAL REPORTS. A Superharmonic Prior for the Autoregressive Process of the Second Order
MATHEMATICAL ENGINEERING TECHNICAL REPORTS A Superharmonic Prior for the Autoregressive Process of the Second Order Fuyuhiko TANAKA and Fumiyasu KOMAKI METR 2006 30 May 2006 DEPARTMENT OF MATHEMATICAL
More informationGSI Geometric Science of Information, Paris, August Dimensionality reduction for classification of stochastic fibre radiographs
GSI2013 - Geometric Science of Information, Paris, 28-30 August 2013 Dimensionality reduction for classification of stochastic fibre radiographs C.T.J. Dodson 1 and W.W. Sampson 2 School of Mathematics
More informationx. Figure 1: Examples of univariate Gaussian pdfs N (x; µ, σ 2 ).
.8.6 µ =, σ = 1 µ = 1, σ = 1 / µ =, σ =.. 3 1 1 3 x Figure 1: Examples of univariate Gaussian pdfs N (x; µ, σ ). The Gaussian distribution Probably the most-important distribution in all of statistics
More informationDually Flat Geometries in the State Space of Statistical Models
1/ 12 Dually Flat Geometries in the State Space of Statistical Models Jan Naudts Universiteit Antwerpen ECEA, November 2016 J. Naudts, Dually Flat Geometries in the State Space of Statistical Models. In
More informationCopula Regression RAHUL A. PARSA DRAKE UNIVERSITY & STUART A. KLUGMAN SOCIETY OF ACTUARIES CASUALTY ACTUARIAL SOCIETY MAY 18,2011
Copula Regression RAHUL A. PARSA DRAKE UNIVERSITY & STUART A. KLUGMAN SOCIETY OF ACTUARIES CASUALTY ACTUARIAL SOCIETY MAY 18,2011 Outline Ordinary Least Squares (OLS) Regression Generalized Linear Models
More information4.7 The Levi-Civita connection and parallel transport
Classnotes: Geometry & Control of Dynamical Systems, M. Kawski. April 21, 2009 138 4.7 The Levi-Civita connection and parallel transport In the earlier investigation, characterizing the shortest curves
More informationSTATISTICAL PROPERTIES FOR THE GENERALIZED COMPOUND GAMMA DISTRIBUTION
STATISTICAL PROPERTIES FOR THE GENERALIZED COMPOUND GAMMA DISTRIBUTION SY Mead, M. L' Adratt In this paper, we introduce a general form of compound gamma distriution with four parameters ( or a generalized
More informationHolonomy groups. Thomas Leistner. Mathematics Colloquium School of Mathematics and Physics The University of Queensland. October 31, 2011 May 28, 2012
Holonomy groups Thomas Leistner Mathematics Colloquium School of Mathematics and Physics The University of Queensland October 31, 2011 May 28, 2012 1/17 The notion of holonomy groups is based on Parallel
More informationTHE GEOMETRY OF B-FIELDS. Nigel Hitchin (Oxford) Odense November 26th 2009
THE GEOMETRY OF B-FIELDS Nigel Hitchin (Oxford) Odense November 26th 2009 THE B-FIELD IN PHYSICS B = i,j B ij dx i dx j flux: db = H a closed three-form Born-Infeld action: det(g ij + B ij ) complexified
More informationSection 2. Basic formulas and identities in Riemannian geometry
Section 2. Basic formulas and identities in Riemannian geometry Weimin Sheng and 1. Bianchi identities The first and second Bianchi identities are R ijkl + R iklj + R iljk = 0 R ijkl,m + R ijlm,k + R ijmk,l
More informationUniversity of Lisbon, Portugal
Development and comparative study of two near-exact approximations to the distribution of the product of an odd number of independent Beta random variables Luís M. Grilo a,, Carlos A. Coelho b, a Dep.
More informationA NOTE ON FISCHER-MARSDEN S CONJECTURE
PROCEEDINGS OF THE AMERICAN MATHEMATICAL SOCIETY Volume 125, Number 3, March 1997, Pages 901 905 S 0002-9939(97)03635-6 A NOTE ON FISCHER-MARSDEN S CONJECTURE YING SHEN (Communicated by Peter Li) Abstract.
More informationHyperbolic Geometric Flow
Hyperbolic Geometric Flow Kefeng Liu Zhejiang University UCLA Page 1 of 41 Outline Introduction Hyperbolic geometric flow Local existence and nonlinear stability Wave character of metrics and curvatures
More informationUniform Correlation Mixture of Bivariate Normal Distributions and. Hypercubically-contoured Densities That Are Marginally Normal
Uniform Correlation Mixture of Bivariate Normal Distributions and Hypercubically-contoured Densities That Are Marginally Normal Kai Zhang Department of Statistics and Operations Research University of
More informationL 2 Geometry of the Symplectomorphism Group
University of Notre Dame Workshop on Innite Dimensional Geometry, Vienna 2015 Outline 1 The Exponential Map on D s ω(m) 2 Existence of Multiplicity of Outline 1 The Exponential Map on D s ω(m) 2 Existence
More informationAn Introduction to Riemann-Finsler Geometry
D. Bao S.-S. Chern Z. Shen An Introduction to Riemann-Finsler Geometry With 20 Illustrations Springer Contents Preface Acknowledgments vn xiii PART ONE Finsler Manifolds and Their Curvature CHAPTER 1 Finsler
More informationINVERSE CASCADE on HYPERBOLIC PLANE
INVERSE CASCADE on HYPERBOLIC PLANE Krzysztof Gawe,dzki based on work in progress with Grisha Falkovich Nordita, October 2011 to Uriel No road is long with good company turkish proverb V.I. Arnold, Ann.
More informationDimensionality Reduction for Information Geometric Characterization of Surface Topographies. Dodson, CTJ and Mettanen, M and Sampson, WW
Dimensionality Reduction for Information Geometric Characterization of Surface Topographies Dodson, CTJ and Mettanen, M and Sampson, WW 2017 MIMS EPrint: 2016.58 Manchester Institute for Mathematical Sciences
More informationManifold Monte Carlo Methods
Manifold Monte Carlo Methods Mark Girolami Department of Statistical Science University College London Joint work with Ben Calderhead Research Section Ordinary Meeting The Royal Statistical Society October
More informationLECTURE 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 ( V ) the set of 4-tensors that is anti-symmetric with respect to
More informationLecture No 1 Introduction to Diffusion equations The heat equat
Lecture No 1 Introduction to Diffusion equations The heat equation Columbia University IAS summer program June, 2009 Outline of the lectures We will discuss some basic models of diffusion equations and
More informationarxiv:math/ v1 [math.dg] 19 Nov 2004
arxiv:math/04426v [math.dg] 9 Nov 2004 REMARKS ON GRADIENT RICCI SOLITONS LI MA Abstract. In this paper, we study the gradient Ricci soliton equation on a complete Riemannian manifold. We show that under
More informationProbability on a Riemannian Manifold
Probability on a Riemannian Manifold Jennifer Pajda-De La O December 2, 2015 1 Introduction We discuss how we can construct probability theory on a Riemannian manifold. We make comparisons to this and
More informationB-FIELDS, GERBES AND GENERALIZED GEOMETRY
B-FIELDS, GERBES AND GENERALIZED GEOMETRY Nigel Hitchin (Oxford) Durham Symposium July 29th 2005 1 PART 1 (i) BACKGROUND (ii) GERBES (iii) GENERALIZED COMPLEX STRUCTURES (iv) GENERALIZED KÄHLER STRUCTURES
More informationRandom Matrix Eigenvalue Problems in Probabilistic Structural Mechanics
Random Matrix Eigenvalue Problems in Probabilistic Structural Mechanics S Adhikari Department of Aerospace Engineering, University of Bristol, Bristol, U.K. URL: http://www.aer.bris.ac.uk/contact/academic/adhikari/home.html
More informationHodge de Rham decomposition for an L 2 space of differfential 2-forms on path spaces
Hodge de Rham decomposition for an L 2 space of differfential 2-forms on path spaces K. D. Elworthy and Xue-Mei Li For a compact Riemannian manifold the space L 2 A of L 2 differential forms decomposes
More informationHYPERSURFACES OF EUCLIDEAN SPACE AS GRADIENT RICCI SOLITONS *
ANALELE ŞTIINŢIFICE ALE UNIVERSITĂŢII AL.I. CUZA DIN IAŞI (S.N.) MATEMATICĂ, Tomul LXI, 2015, f.2 HYPERSURFACES OF EUCLIDEAN SPACE AS GRADIENT RICCI SOLITONS * BY HANA AL-SODAIS, HAILA ALODAN and SHARIEF
More informationLecture 1: August 28
36-705: Intermediate Statistics Fall 2017 Lecturer: Siva Balakrishnan Lecture 1: August 28 Our broad goal for the first few lectures is to try to understand the behaviour of sums of independent random
More informationStat 5101 Lecture Notes
Stat 5101 Lecture Notes Charles J. Geyer Copyright 1998, 1999, 2000, 2001 by Charles J. Geyer May 7, 2001 ii Stat 5101 (Geyer) Course Notes Contents 1 Random Variables and Change of Variables 1 1.1 Random
More informationLegendre surfaces whose mean curvature vectors are eigenvectors of the Laplace operator
Note di Matematica 22, n. 1, 2003, 9 58. Legendre surfaces whose mean curvature vectors are eigenvectors of the Laplace operator Tooru Sasahara Department of Mathematics, Hokkaido University, Sapporo 060-0810,
More information1. Geometry of the unit tangent bundle
1 1. Geometry of the unit tangent bundle The main reference for this section is [8]. In the following, we consider (M, g) an n-dimensional smooth manifold endowed with a Riemannian metric g. 1.1. Notations
More informationTensors, and differential forms - Lecture 2
Tensors, and differential forms - Lecture 2 1 Introduction The concept of a tensor is derived from considering the properties of a function under a transformation of the coordinate system. A description
More informationContinuous Univariate Distributions
Continuous Univariate Distributions Volume 1 Second Edition NORMAN L. JOHNSON University of North Carolina Chapel Hill, North Carolina SAMUEL KOTZ University of Maryland College Park, Maryland N. BALAKRISHNAN
More informationSOME ALMOST COMPLEX STRUCTURES WITH NORDEN METRIC ON THE TANGENT BUNDLE OF A SPACE FORM
ANALELE ŞTIINŢIFICE ALE UNIVERSITĂŢII AL.I.CUZA IAŞI Tomul XLVI, s.i a, Matematică, 2000, f.1. SOME ALMOST COMPLEX STRUCTURES WITH NORDEN METRIC ON THE TANGENT BUNDLE OF A SPACE FORM BY N. PAPAGHIUC 1
More informationContinuous Univariate Distributions
Continuous Univariate Distributions Volume 2 Second Edition NORMAN L. JOHNSON University of North Carolina Chapel Hill, North Carolina SAMUEL KOTZ University of Maryland College Park, Maryland N. BALAKRISHNAN
More informationModelling Dropouts by Conditional Distribution, a Copula-Based Approach
The 8th Tartu Conference on MULTIVARIATE STATISTICS, The 6th Conference on MULTIVARIATE DISTRIBUTIONS with Fixed Marginals Modelling Dropouts by Conditional Distribution, a Copula-Based Approach Ene Käärik
More informationHyperbolic Geometric Flow
Hyperbolic Geometric Flow Kefeng Liu CMS and UCLA August 20, 2007, Dong Conference Page 1 of 51 Outline: Joint works with D. Kong and W. Dai Motivation Hyperbolic geometric flow Local existence and nonlinear
More information1 First and second variational formulas for area
1 First and second variational formulas for area In this chapter, we will derive the first and second variational formulas for the area of a submanifold. This will be useful in our later discussion on
More informationNewman-Penrose formalism in higher dimensions
Newman-Penrose formalism in higher dimensions V. Pravda various parts in collaboration with: A. Coley, R. Milson, M. Ortaggio and A. Pravdová Introduction - algebraic classification in four dimensions
More informationAnn. Acad. Rom. Sci. Ser. Math. Appl. Vol. 8, No. 1/2016. The space l (X) Namita Das
ISSN 2066-6594 Ann. Acad. Rom. Sci. Ser. Math. Appl. Vol. 8, No. 1/2016 The space l (X) Namita Das Astract In this paper we have shown that the space c n n 0 cannot e complemented in l n n and c 0 (H)
More informationON THE FOLIATION OF SPACE-TIME BY CONSTANT MEAN CURVATURE HYPERSURFACES
ON THE FOLIATION OF SPACE-TIME BY CONSTANT MEAN CURVATURE HYPERSURFACES CLAUS GERHARDT Abstract. We prove that the mean curvature τ of the slices given by a constant mean curvature foliation can be used
More informationThe Riemann curvature tensor, its invariants, and their use in the classification of spacetimes
DigitalCommons@USU Presentations and Publications 3-20-2015 The Riemann curvature tensor, its invariants, and their use in the classification of spacetimes Follow this and additional works at: http://digitalcommons.usu.edu/dg_pres
More informationOn the existence of isoperimetric extremals of rotation and the fundamental equations of rotary diffeomorphisms
XV II th International Conference of Geometry, Integrability and Quantization Sveti Konstantin i Elena 2016 On the existence of isoperimetric extremals of rotation and the fundamental equations of rotary
More informationOn a simple construction of bivariate probability functions with fixed marginals 1
On a simple construction of bivariate probability functions with fixed marginals 1 Djilali AIT AOUDIA a, Éric MARCHANDb,2 a Université du Québec à Montréal, Département de mathématiques, 201, Ave Président-Kennedy
More informationBERGMAN KERNEL ON COMPACT KÄHLER MANIFOLDS
BERGMAN KERNEL ON COMPACT KÄHLER MANIFOLDS SHOO SETO Abstract. These are the notes to an expository talk I plan to give at MGSC on Kähler Geometry aimed for beginning graduate students in hopes to motivate
More informationAlgebraic Information Geometry for Learning Machines with Singularities
Algebraic Information Geometry for Learning Machines with Singularities Sumio Watanabe Precision and Intelligence Laboratory Tokyo Institute of Technology 4259 Nagatsuta, Midori-ku, Yokohama, 226-8503
More informationManifolds in Fluid Dynamics
Manifolds in Fluid Dynamics Justin Ryan 25 April 2011 1 Preliminary Remarks In studying fluid dynamics it is useful to employ two different perspectives of a fluid flowing through a domain D. The Eulerian
More informationHolonomy groups. Thomas Leistner. School of Mathematical Sciences Colloquium University of Adelaide, May 7, /15
Holonomy groups Thomas Leistner School of Mathematical Sciences Colloquium University of Adelaide, May 7, 2010 1/15 The notion of holonomy groups is based on Parallel translation Let γ : [0, 1] R 2 be
More informationThe continuum limit of distributed dislocations
The continuum limit of distriuted dislocations Cy Maor! Institute of Mathematics, Herew University!! Conference on non-linearity, transport, physics, and patterns! Fields Institute, Octoer 2014 Different
More informationNatural Gradient for the Gaussian Distribution via Least Squares Regression
Natural Gradient for the Gaussian Distribution via Least Squares Regression Luigi Malagò Department of Electrical and Electronic Engineering Shinshu University & INRIA Saclay Île-de-France 4-17-1 Wakasato,
More informationUNIVERSITY OF DUBLIN
UNIVERSITY OF DUBLIN TRINITY COLLEGE JS & SS Mathematics SS Theoretical Physics SS TSM Mathematics Faculty of Engineering, Mathematics and Science school of mathematics Trinity Term 2015 Module MA3429
More informationPAPER 52 GENERAL RELATIVITY
MATHEMATICAL TRIPOS Part III Monday, 1 June, 2015 9:00 am to 12:00 pm PAPER 52 GENERAL RELATIVITY Attempt no more than THREE questions. There are FOUR questions in total. The questions carry equal weight.
More informationON VARIATION OF THE METRIC TENSOR IN THE ACTION OF A PHYSICAL FIELD
ON VARIATION OF THE METRIC TENSOR IN THE ACTION OF A PHYSICAL FIELD L.D. Raigorodski Abstract The application of the variations of the metric tensor in action integrals of physical fields is examined.
More information