Ordre segmentSegment parentTitreComplément titreAuteursMots-clésAbstractURL documentURL VidéosDOI
1Opening Session (chaired by Frédéric Barbaresco)SEE-GSI'15 Opening sessionFrédéric Barbarescohttp://www.see.asso.fr/en/fichier/14256_see-gsi-15-opening-session
2Keynote speach Matilde Marcolli (chaired by Daniel Bennequin)From Geometry and Physics to Computational LinguisticsMatilde Marcolli
I will show how techniques from geometry (algebraic geometry and topology) and physics (statistical physics) can be applied to Linguistics, in order to provide a computational approach to questions of syntactic 
3Random Geometry/Homology (chaired by Laurent Decreusefond/Frédéric Chazal)The extremal index for a random tessellationNicolas ChenavierExtreme values, Poisson point process, Random tessellations
Let m be a random tessellation in R d , d ≥ 1, observed in the window W p = ρ1/d[0, 1] d , ρ > 0, and let f be a geometrical characteristic. We investigate the asymptotic behaviour of the maximum of f(C) over all cells C ∈ m with nucleus W p as ρ goes to infinity.When the normalized maximum converges, we show that its asymptotic distribution depends on the so-called extremal index. Two examples of extremal indices are provided for Poisson-Voronoi and Poisson-Delaunay tessellations.
3Random Geometry/Homology (chaired by Laurent Decreusefond/Frédéric Chazal)A two-color interacting random balls model for co-localization analysis of proteinsCharles Kervrann, Frederic Lavancier
A model of two-type (or two-color) interacting random balls is introduced. Each colored random set is a union of random balls and the interaction relies on the volume of the intersection between the two random sets. This model is motivated by the detection and quantification of co-localization between two proteins. Simulation and inference are discussed. Since all individual balls cannot been identified, e.g. a ball may contain another one, standard methods of inference as likelihood or pseudolikelihood are not available and we apply the Takacs-Fiksel method with a specific choice of test functions.
3Random Geometry/Homology (chaired by Laurent Decreusefond/Frédéric Chazal)Asymptotics of superposition of point processesAurélien Vasseur, Laurent DecreusefondGinibre point process, Poisson point process, Stein’s method, Stochastic geometry, β-Ginibre point process
The characteristic independence property of Poisson point processes gives an intuitive way to explain why a sequence of point processes becoming less and less repulsive can converge to a Poisson point process. The aim of this paper is to show this convergence for sequences built by superposing, thinning or rescaling determinantal processes. We use Papangelou intensities and Stein’s method to prove this result with a topology based on total variation distance.
3Random Geometry/Homology (chaired by Laurent Decreusefond/Frédéric Chazal)Asymptotic properties of random polytopesPierre Calka
Random polytopes have constituted some of the central objects of stochastic geometry for more than 150 years. They are in general generated as convex hulls of a random set of points in the Euclidean space. The study of such models requires the use of ingredients coming from both convex geometry and probability theory. In the last decades, the study has been focused on their asymptotic properties and in particular expectation and variance estimates. In several joint works with Tomasz Schreiber and J. E. Yukich, we have investigated the scaling limit of several models (uniform model in the unit-ball, uniform model in a smooth convex body, Gaussian model) and have deduced from it limiting variances for several geometric characteristics including the number of k-dimensional faces and the volume. In this paper, we survey the most recent advances on these questions and we emphasize the particular cases of random polytopes in the unit-ball and Gaussian polytopes.
3Random Geometry/Homology (chaired by Laurent Decreusefond/Frédéric Chazal)Asymmetric Topologies on Statistical ManifoldsRoman Belavkin
Asymmetric information distances are used to define asymmetric norms and quasimetrics on the statistical manifold and its dual space of random variables. Quasimetric topology, generated by the Kullback-Leibler (KL) divergence, is considered as the main example, and some of its topological properties are investigated.
4Computational Information Geometry (chaired by Frank Nielsen, Paul Marriott)Geometry of Goodness-of-Fit Testing in High Dimensional Low Sample Size ModellingFrank Critchley, Germain Van Bever, Paul Marriott, Radka Sabolova
We introduce a new approach to goodness-of-fit testing in the high dimensional, sparse extended multinomial context. The paper takes a computational information geometric approach, extending classical higher order asymptotic theory. We show why the Wald – equivalently, the Pearson X2 and score statistics – are unworkable in this context, but that the deviance has a simple, accurate and tractable sampling distribution even for moderate sample sizes. Issues of uniformity of asymptotic approximations across model space are discussed. A variety of important applications and extensions are noted.
4Computational Information Geometry (chaired by Frank Nielsen, Paul Marriott)Computing Boundaries in Local Mixture ModelsPaul Marriott, Vahed MaroufyComputational information geometry, Computing boundaries, Embedded manifolds, Local mixture models, Polytopes, Ruled and developable surfaces
Local mixture models give an inferentially tractable but still flexible alternative to general mixture models. Their parameter space naturally includes boundaries; near these the behaviour of the likelihood is not standard. This paper shows how convex and differential geometries help in characterising these boundaries. In particular the geometry of polytopes, ruled and developable surfaces is exploited to develop efficient inferential algorithms.
4Computational Information Geometry (chaired by Frank Nielsen, Paul Marriott)Approximating Covering and Minimum Enclosing Balls in Hyperbolic GeometryFrank Nielsen, Gaëtan Hadjeres
We generalize the O(dnϵ2)-time (1 + ε)-approximation algorithm for the smallest enclosing Euclidean ball [2,10] to point sets in hyperbolic geometry of arbitrary dimension. We guarantee a O(1/ϵ2) convergence time by using a closed-form formula to compute the geodesic α-midpoint between any two points. Those results allow us to apply the hyperbolic k-center clustering for statistical location-scale families or for multivariate spherical normal distributions by using their Fisher information matrix as the underlying Riemannian hyperbolic metric.
4Computational Information Geometry (chaired by Frank Nielsen, Paul Marriott)From Euclidean to Riemannian Means Information Geometry for SSVEP ClassificationEmmanuel Kalunga, Eric Monacelli, Karim Djouani, Quentin Barthélemy, Sylvain Chevallier, Yskandar HamamBrain- Computer, Information geometry, Interfaces, Riemannian means, Steady State, Visually Evoked Potentials
Brain Computer Interfaces (BCI) based on electroencephalography (EEG) rely on multichannel brain signal processing. Most of the state-of-the-art approaches deal with covariance matrices, and indeed Riemannian geometry has provided a substantial framework for developing new algorithms. Most notably, a straightforward algorithm such as Minimum Distance to Mean yields competitive results when applied with a Riemannian distance. This applicative contribution aims at assessing the impact of several distances on real EEG dataset, as the invariances embedded in those distances have an influence on the classification accuracy. Euclidean and Riemannian distances and means are compared both in term of quality of results and of computational load.
4Computational Information Geometry (chaired by Frank Nielsen, Paul Marriott)Group Theoretical Study on Geodesics for the Elliptical ModelsHiroto Inoue
We consider the geodesic equation on the elliptical model, which is a generalization of the normal model. More precisely, we characterize this manifold from the group theoretical view point and formulate Eriksen’s procedure to obtain geodesics on normal model and give an alternative proof for it.
4Computational Information Geometry (chaired by Frank Nielsen, Paul Marriott)Path connectedness on a space of probability density functionsOsamu Komori, Shinto Eguchi
We introduce a class of paths or one-parameter models connecting arbitrary two probability density functions (pdf’s). The class is derived by employing the Kolmogorov-Nagumo average between the two pdf’s. There is a variety of such path connectedness on the space of pdf’s since the Kolmogorov-Nagumo average is applicable for any convex and strictly increasing function. The information geometric insight is provided for understanding probabilistic properties for statistical methods associated with the path connectedness. The one-parameter model is extended to a multidimensional model, on which the statistical inference is characterized by sufficient statistics.
4Computational Information Geometry (chaired by Frank Nielsen, Paul Marriott)Computational Information Geometry: mixture modellingFrank Critchley, Germain Van Bever, Paul Marriott, Radka Sabolovahttp://www.see.asso.fr/en/fichier/14268_computational-information-geometry-mixture-modellinghttps://www.youtube.com/watch?v=lpzCTiW1QHI
5Bayesian and Information Geometry for Inverse Problems (chaired by Ali Mohammad-Djafari, Olivier Swander)Stochastic PDE projection on manifolds Assumed-Density and Galerkin FiltersDamiano Brigo, John Armstrong
We review the manifold projection method for stochastic nonlinear filtering in a more general setting than in our previous paper in Geometric Science of Information 2013. We still use a Hilbert space structure on a space of probability densities to project the infinite dimensional stochastic partial differential equation for the optimal filter onto a finite dimensional exponential or mixture family, respectively, with two different metrics, the Hellinger distance and the L2 direct metric. This reduces the problem to finite dimensional stochastic differential equations. In this paper we summarize a previous equivalence result between Assumed Density Filters (ADF) and Hellinger/Exponential projection filters, and introduce a new equivalence between Galerkin method based filters and Direct metric/Mixture projection filters. This result allows us to give a rigorous geometric interpretation to ADF and Galerkin filters. We also discuss the different finite-dimensional filters obtained when projecting the stochastic partial differential equation for either the normalized (Kushner-Stratonovich) or a specific unnormalized (Zakai) density of the optimal filter.
5Bayesian and Information Geometry for Inverse Problems (chaired by Ali Mohammad-Djafari, Olivier Swander)Variational Bayesian Approximation method for Classification and Clustering with a mixture of StudenAli Mohammad-Djafari
Clustering, classification and Pattern Recognition in a set of data are between the most important tasks in statistical researches and in many applications. In this paper, we propose to use a mixture of Student-t distribution model for the data via a hierarchical graphical model and the Bayesian framework to do these tasks. The main advantages of this model is that the model accounts for the uncertainties of variances and covariances and we can use the Variational Bayesian Approximation (VBA) methods to obtain fast algorithms to be able to handle large data sets.
5Bayesian and Information Geometry for Inverse Problems (chaired by Ali Mohammad-Djafari, Olivier Swander)Differential geometric properties of textile plotTomonari Sei, Ushio Tanaka
The textile plot proposed by Kumasaka and Shibata (2008) is a method for data visualization. The method transforms a data matrix in order to draw a parallel coordinate plot. In this paper, we investigate a set of matrices induced by the textile plot, which we call the textile set, from a geometrical viewpoint. It is shown that the textile set is written as the union of two differentiable manifolds if data matrices are restricted to be full-rank.
5Bayesian and Information Geometry for Inverse Problems (chaired by Ali Mohammad-Djafari, Olivier Swander)A generalization of independence and multivariate Student's t-distributionsHiroshi Matsuzoe, Monta Sakamoto
In anomalous statistical physics, deformed algebraic structures are important objects. Heavily tailed probability distributions, such as Student’s t-distributions, are characterized by deformed algebras. In addition, deformed algebras cause deformations of expectations and independences of random variables. Hence, a generalization of independence for multivariate Student’s t-distribution is studied in this paper. Even if two random variables which follow to univariate Student’s t-distributions are independent, the joint probability distribution of these two distributions is not a bivariate Student’s t-distribution. It is shown that a bivariate Student’s t-distribution is obtained from two univariate Student’s t-distributions under q-deformed independence.
6Hessian Information Geometry (chaired by Shun-Ichi Amari, Michel Boyom)New metric and connections in statistical manifoldsCharles Cavalcante, David de Souza, Rui Vigelis
We define a metric and a family of α-connections in statistical manifolds, based on ϕ-divergence, which emerges in the framework of ϕ-families of probability distributions. This metric and α-connections generalize the Fisher information metric and Amari’s α-connections. We also investigate the parallel transport associated with the α-connection for α = 1.
6Hessian Information Geometry (chaired by Shun-Ichi Amari, Michel Boyom)Curvatures of Statistical StructuresBarbara OpozdaAffine connection, Curvature tensor, Laplacian Bochner’s technique, Ricci tensor, Sectional curvature
Curvature properties for statistical structures are studied. The study deals with the curvature tensor of statistical connections and their duals as well as the Ricci tensor of the connections, Laplacians and the curvature operator. Two concepts of sectional curvature are introduced. The meaning of the notions is illustrated by presenting few exemplary theorems.
6Hessian Information Geometry (chaired by Shun-Ichi Amari, Michel Boyom)The Pontryagin Forms of Hessian ManifoldsJohn Armstrong, Shun-Ichi Amari
We show that Hessian manifolds of dimensions 4 and above must have vanishing Pontryagin forms. This gives a topological obstruction to the existence of Hessian metrics. We find an additional explicit curvature identity for Hessian 4-manifolds. By contrast, we show that all analytic Riemannian 2-manifolds are Hessian.
6Hessian Information Geometry (chaired by Shun-Ichi Amari, Michel Boyom)Matrix realization of a homogeneous coneHideyuki IshiHessian metric, Homogeneous cone, Left-symmetric algebra
Based on the theory of compact normal left-symmetric algebra (clan), we realize every homogeneous cone as a set of positive definite real symmetric matrices, where homogeneous Hessian metrics as well as a transitive group action on the cone are described efficiently.
6Hessian Information Geometry (chaired by Shun-Ichi Amari, Michel Boyom)Multiply CR-Warped Product Statistical Submanifolds of a Holomorphic Stastistical Space FormJamali Mohammed, Michel Boyom, Shahid Hasan
In this article, we derive an inequality satisfied by the squared norm of the imbedding curvature tensor of Multiply CR-warped product statistical submanifolds N of holomorphic statistical space forms M. Furthermore, we prove that under certain geometric conditions, N and M become Einstein.
7Topological forms and Information (chaired by Daniel Bennequin, Pierre Baudot)Information Algebras and their ApplicationsMatilde Marcolli
In this lecture we will present joint work with Ryan Thorngren on thermodynamic semirings and entropy operads, with Nicolas Tedeschi on Birkhoff factorization in thermodynamic semirings, ongoing work with Marcus Bintz on tropicalization of Feynman graph hypersurfaces and Potts model hypersurfaces, and their thermodynamic deformations, and ongoing work by the author on applications of thermodynamic semirings to models of morphology and syntax in Computational Linguistics.
7Topological forms and Information (chaired by Daniel Bennequin, Pierre Baudot)Finite polylogarithms, their multiple analogues and the Shannon entropyHerbert Gangl, Philippe Elbaz-Vincent
We show that the entropy function–and hence the finite 1-logarithm–behaves a lot like certain derivations. We recall its cohomological interpretation as a 2-cocycle and also deduce 2n-cocycles for any n. Finally, we give some identities for finite multiple polylogarithms together with number theoretic applications.
7Topological forms and Information (chaired by Daniel Bennequin, Pierre Baudot)Heights of toric varieties, integration over polytopes and entropyJosé Ignacio Burgos Gil, Martin Sombra, Patrice Philippon
We present a dictionary between arithmetic geometry of toric varieties and convex analysis. This correspondence allows for effective computations of arithmetic invariants of these varieties. In particular, combined with a closed formula for the integration of a class of functions over polytopes, it gives a number of new values for the height (arithmetic analog of the degree) of toric varieties, with respect to interesting metrics arising from polytopes. In some cases these heights are interpreted as the average entropy of a family of random processes.
7Topological forms and Information (chaired by Daniel Bennequin, Pierre Baudot)Characterization and Estimation of the Variations of a Random Convex Set by it's Mean $n$-Variogram : Application to the Boolean ModelJean-Charles Pinoli, Johan Debayle, Saïd RahmaniBoolean model, Geometric covariogram, Mixed volumes, n points set probability, Particle size distribution, Random set Shape variations
In this paper we propose a method to characterize and estimate the variations of a random convex set Ξ0 in terms of shape, size and direction. The mean n-variogram γ(n)Ξ0:(u1⋯un)↦E[νd(Ξ0∩(Ξ0−u1)⋯∩(Ξ0−un))] of a random convex set Ξ0 on ℝ d reveals information on the n th order structure of Ξ0. Especially we will show that considering the mean n-variograms of the dilated random sets Ξ0 ⊕ rK by an homothetic convex family rKr > 0, it’s possible to estimate some characteristic of the n th order structure of Ξ0. If we make a judicious choice of K, it provides relevant measures of Ξ0. Fortunately the germ-grain model is stable by convex dilatations, furthermore the mean n-variogram of the primary grain is estimable in several type of stationary germ-grain models by the so called n-points probability function. Here we will only focus on the Boolean model, in the planar case we will show how to estimate the n th order structure of the random vector composed by the mixed volumes t (A(Ξ0),W(Ξ0,K)) of the primary grain, and we will describe a procedure to do it from a realization of the Boolean model in a bounded window. We will prove that this knowledge for all convex body K is sufficient to fully characterize the so called difference body of the grain Ξ0⊕˘Ξ0. we will be discussing the choice of the element K, by choosing a ball, the mixed volumes coincide with the Minkowski’s functional of Ξ0 therefore we obtain the moments of the random vector composed of the area and perimeter t (A(Ξ0),U(Ξ)). By choosing a segment oriented by θ we obtain estimates for the moments of the random vector composed by the area and the Ferret’s diameter in the direction θ, t((A(Ξ0),HΞ0(θ)). Finally, we will evaluate the performance of the method on a Boolean model with rectangular grain for the estimation of the second order moments of the random vectors t (A(Ξ0),U(Ξ0)) and t((A(Ξ0),HΞ0(θ)).
8Short course (chaired by Roger Balian)Geometry on the set of quantum states and quantum correlationsDominique Spehnerhttp://www.see.asso.fr/en/fichier/14277_geometry-set-quantum-states-and-quantum-correlationshttps://www.youtube.com/watch?v=5Nj5afyivI8
11Keynote speach Marc Arnaudon (chaired by Frank Nielsen)Stochastic Euler-Poincaré reductionMarc Arnaudon
We will prove a Euler-Poincaré reduction theorem for stochastic processes taking values in a Lie group, which is a generalization of the Lagrangian version of reduction and its associated variational principles. We will also show examples of its application to the rigid body and to the group of diffeomorphisms, which includes the Navier-Stokes equation on a bounded domain and the Camassa-Holm equation.
12Information Geometry Optimization (chaired by Giovanni Pistone, Yann Ollivier)Laplace's rule of succession in information geometryYann Ollivier
When observing data x1, . . . , x t modelled by a probabilistic distribution pθ(x), the maximum likelihood (ML) estimator θML = arg max θ Σti=1 ln pθ(x i ) cannot, in general, safely be used to predict xt + 1. For instance, for a Bernoulli process, if only “tails” have been observed so far, the probability of “heads” is estimated to 0. (Thus for the standard log-loss scoring rule, this results in infinite loss the first time “heads” appears.)
12Information Geometry Optimization (chaired by Giovanni Pistone, Yann Ollivier)Standard Divergence in Manifold of Dual Affine ConnectionsNihat Ay, Shun-Ichi Amari
A divergence function defines a Riemannian metric G and dually coupled affine connections (∇, ∇  ∗ ) with respect to it in a manifold M. When M is dually flat, a canonical divergence is known, which is uniquely determined from {G, ∇, ∇  ∗ }. We search for a standard divergence for a general non-flat M. It is introduced by the magnitude of the inverse exponential map, where α = -(1/3) connection plays a fundamental role. The standard divergence is different from the canonical divergence.
12Information Geometry Optimization (chaired by Giovanni Pistone, Yann Ollivier)Transformations and Coupling Relations for Affine ConnectionsJames Tao, Jun Zhang
The statistical structure on a manifold M is predicated upon a special kind of coupling between the Riemannian metric g and a torsion-free affine connection ∇ on the TM, such that ∇ g is totally symmetric, forming, by definition, a “Codazzi pair” { ∇ , g}. In this paper, we first investigate various transformations of affine connections, including additive translation (by an arbitrary (1,2)-tensor K), multiplicative perturbation (through an arbitrary invertible operator L on TM), and conjugation (through a non-degenerate two-form h). We then study the Codazzi coupling of ∇ with h and its coupling with L, and the link between these two couplings. We introduce, as special cases of K-translations, various transformations that generalize traditional projective and dual-projective transformations, and study their commutativity with L-perturbation and h-conjugation transformations. Our derivations allow affine connections to carry torsion, and we investigate conditions under which torsions are preserved by the various transformations mentioned above. Our systematic approach establishes a general setting for the study of Information Geometry based on transformations and coupling relations of affine connections – in particular, we provide a generalization of conformal-projective transformation.
12Information Geometry Optimization (chaired by Giovanni Pistone, Yann Ollivier)Online k-MLE for mixture modeling with exponential familiesChristophe Saint-Jean, Frank Nielsenk-MLE Wishart distribution, Mixture modeling, Online learning
This paper address the problem of online learning finite statistical mixtures of exponential families. A short review of the Expectation-Maximization (EM) algorithm and its online extensions is done. From these extensions and the description of the k-Maximum Likelihood Estimator (k-MLE), three online extensions are proposed for this latter. To illustrate them, we consider the case of mixtures of Wishart distributions by giving details and providing some experiments.
12Information Geometry Optimization (chaired by Giovanni Pistone, Yann Ollivier)Second-order Optimization over the Multivariate Gaussian DistributionGiovanni Pistone, Luigi Malagò
We discuss the optimization of the stochastic relaxation of a real-valued function, i.e., we introduce a new search space given by a statistical model and we optimize the expected value of the original function with respect to a distribution in the model. From the point of view of Information Geometry, statistical models are Riemannian manifolds of distributions endowed with the Fisher information metric, thus the stochastic relaxation can be seen as a continuous optimization problem defined over a differentiable manifold. In this paper we explore the second-order geometry of the exponential family, with applications to the multivariate Gaussian distributions, to generalize second-order optimization methods. Besides the Riemannian Hessian, we introduce the exponential and the mixture Hessians, which come from the dually flat structure of an exponential family. This allows us to obtain different Taylor formulæ according to the choice of the Hessian and of the geodesic used, and thus different approaches to the design of second-order methods, such as the Newton method.
12Information Geometry Optimization (chaired by Giovanni Pistone, Yann Ollivier)The information geometry of mirror descentGarvesh Raskutti, Sayan Mukherjee
We prove the equivalence of two online learning algorithms, mirror descent and natural gradient descent. Both mirror descent and natural gradient descent are generalizations of online gradient descent when the parameter of interest lies on a non-Euclidean manifold. Natural gradient descent selects the steepest descent direction along a Riemannian manifold by multiplying the standard gradient by the inverse of the metric tensor. Mirror descent induces non-Euclidean structure by solving iterative optimization problems using different proximity functions. In this paper, we prove that mirror descent induced by a Bregman divergence proximity functions is equivalent to the natural gradient descent algorithm on the Riemannian manifold in the dual coordinate system.We use techniques from convex analysis and connections between Riemannian manifolds, Bregman divergences and convexity to prove this result. This equivalence between natural gradient descent and mirror descent, implies that (1) mirror descent is the steepest descent direction along the Riemannian manifold corresponding to the choice of Bregman divergence and (2) mirror descent with log-likelihood loss applied to parameter estimation in exponential families asymptotically achieves the classical Cramér-Rao lower bound.
13Geometry of Time Series and Linear Dynamical systems (chaired by Bijan Afsari, Arshia Cont)TS-GNPR Clustering Random Walk Time SeriesFrank Nielsen, Gautier Marti, Philippe Donnat, Philippe Very
We present in this paper a novel non-parametric approach useful for clustering independent identically distributed stochastic processes. We introduce a pre-processing step consisting in mapping multivariate independent and identically distributed samples from random variables to a generic non-parametric representation which factorizes dependency and marginal distribution apart without losing any information. An associated metric is defined where the balance between random variables dependency and distribution information is controlled by a single parameter. This mixing parameter can be learned or played with by a practitioner, such use is illustrated on the case of clustering financial time series. Experiments, implementation and results obtained on public financial time series are online on a web portal http://www.datagrapple.com .
13Geometry of Time Series and Linear Dynamical systems (chaired by Bijan Afsari, Arshia Cont)A common symmetrization framework for random linear algorithmsAlain Sarlette
This paper highlights some more examples of maps that follow a recently introduced “symmetrization” structure behind the average consensus algorithm. We review among others some generalized consensus settings and coordinate descent optimization.
13Geometry of Time Series and Linear Dynamical systems (chaired by Bijan Afsari, Arshia Cont)New model search for nonlinear recursive models, regressions and autoregressionsAnna-Lena Kißlinger, Wolfgang Stummer3D score surface, AR SARIMA NARX, Autorecursions, Model selection, Nonlinear regression, Scaled Bregman distances
Scaled Bregman distances SBD have turned out to be useful tools for simultaneous estimation and goodness-of-fit-testing in parametric models of random data (streams, clouds). We show how SBD can additionally be used for model preselection (structure detection), i.e. for finding appropriate candidates of model (sub)classes in order to support a desired decision under uncertainty. For this, we exemplarily concentrate on the context of nonlinear recursive models with additional exogenous inputs; as special cases we include nonlinear regressions, linear autoregressive models (e.g. AR, ARIMA, SARIMA time series), and nonlinear autoregressive models with exogenous inputs (NARX). In particular, we outline a corresponding information-geometric 3D computer-graphical selection procedure. Some sample-size asymptotics is given as well.
13Geometry of Time Series and Linear Dynamical systems (chaired by Bijan Afsari, Arshia Cont)Random Pairwise Gossip on CAT(k) Metric SpacesAnass Bellachehab, Jérémie Jakubowicz
In the context of sensor networks, gossip algorithms are a popular, well established technique, for achieving consensus when sensor data are encoded in linear spaces. Gossip algorithms also have several extensions to non linear data spaces. Most of these extensions deal with Riemannian manifolds and use Riemannian gradient descent. This paper, instead, studies gossip in a broader CAT(k) metric setting, encompassing, but not restricted to, several interesting cases of Riemannian manifolds. As it turns out, convergence can be guaranteed as soon as the data lie in a small enough ball of a mere CAT(k) metric space. We also study convergence speed in this setting and establish linear rates of convergence.
14Optimal Transport (chaired by Jean-François Marcotorchino, Alfred Galichon)The nonlinear Bernstein-Schrodinger equation in EconomicsAlfred Galichon, Scott Kominers, Simon Weber
In this paper we relate the Equilibrium Assignment Problem (EAP), which is underlying in several economics models, to a system of nonlinear equations that we call the “nonlinear Bernstein-Schrödinger system”, which is well-known in the linear case, but whose nonlinear extension does not seem to have been studied. We apply this connection to derive an existence result for the EAP, and an efficient computational method.
14Optimal Transport (chaired by Jean-François Marcotorchino, Alfred Galichon)Some geometric consequences of the Schrödinger problemChristian LeonardDisplacement interpolations, Entropic interpolations, Lott-Sturm-Villani theory, Lower bounded curvature of metric spaces, Optimal transport, Schrödinger problem
This note presents a short review of the Schrödinger problem and of the first steps that might lead to interesting consequences in terms of geometry. We stress the analogies between this entropy minimization problem and the renowned optimal transport problem, in search for a theory of lower bounded curvature for metric spaces, including discrete graphs.
14Optimal Transport (chaired by Jean-François Marcotorchino, Alfred Galichon)Optimal Transport, Independance versus Indetermination duality, impact on a new Copula DesignBenoit Huyot, Jean-François Marcotorchino, Yves MabialaCondorcet and relational analysis, Copula theory, Indetermination and independance structures, MKP problem, Optimal transport
This article leans on some previous results already presented in [10], based on the Fréchet’s works,Wilson’s entropy and Minimal Trade models in connectionwith theMKPtransportation problem (MKP, stands for Monge-Kantorovich Problem). Using the duality between “independance” and “indetermination” structures, shown in this former paper, we are in a position to derive a novel approach to design a copula, suitable and efficient for anomaly detection in IT systems analysis.
14Optimal Transport (chaired by Jean-François Marcotorchino, Alfred Galichon)Optimal mass transport over bridgesMichele Pavon, Tryphon Georgiou, Yonxin Chen
We present an overview of our recent work on implementable solutions to the Schrödinger bridge problem and their potential application to optimal transport and various generalizations.
15Information Geometry in Image Analysis (chaired by Yannick Berthoumieu, Geert Verdoolaege)Texture classification using Rao's distance on the space of covariance matricesLionel Bombrun, Salem Said, Yannick BerthoumieuEM algorithm, Information geometry, Mixture estimation, Riemannian centre of mass, Texture classification
The current paper introduces new prior distributions on the zero-mean multivariate Gaussian model, with the aim of applying them to the classification of covariance matrices populations. These new prior distributions are entirely based on the Riemannian geometry of the multivariate Gaussian model. More precisely, the proposed Riemannian Gaussian distribution has two parameters, the centre of mass ˉY and the dispersion parameter σ. Its density with respect to Riemannian volume is proportional to exp(−d2(Y;ˉY)), where d2(Y;ˉY) is the square of Rao’s Riemannian distance. We derive its maximum likelihood estimators and propose an experiment on the VisTex database for the classification of texture images.
15Information Geometry in Image Analysis (chaired by Yannick Berthoumieu, Geert Verdoolaege)Color Texture Discrimination using the Principal Geodesic Distance on a Multivariate Generalized GauAqsa Shabbir, Geert VerdoolaegePrincipal geodesic analysis, Rao geodesic distance, Texture classification
We present a new texture discrimination method for textured color images in the wavelet domain. In each wavelet subband, the correlation between the color bands is modeled by a multivariate generalized Gaussian distribution with fixed shape parameter (Gaussian, Laplacian). On the corresponding Riemannian manifold, the shape of texture clusters is characterized by means of principal geodesic analysis, specifically by the principal geodesic along which the cluster exhibits its largest variance. Then, the similarity of a texture to a class is defined in terms of the Rao geodesic distance on the manifold from the texture’s distribution to its projection on the principal geodesic of that class. This similarity measure is used in a classification scheme, referred to as principal geodesic classification (PGC). It is shown to perform significantly better than several other classifiers.
15Information Geometry in Image Analysis (chaired by Yannick Berthoumieu, Geert Verdoolaege)Bag-of-components an online algorithm for batch learning of mixture modelsFrank Nielsen, Olivier Schwander
Practical estimation of mixture models may be problematic when a large number of observations are involved: for such cases, online versions of Expectation-Maximization may be preferred, avoiding the need to store all the observations before running the algorithms. We introduce a new online method well-suited when both the number of observations is large and lots of mixture models need to be learned from different sets of points. Inspired by dictionary methods, our algorithm begins with a training step which is used to build a dictionary of components. The next step, which can be done online, amounts to populating the weights of the components given each arriving observation. The usage of the dictionary of components shows all its interest when lots of mixtures need to be learned using the same dictionary in order to maximize the return on investment of the training step. We evaluate the proposed method on an artificial dataset built from random Gaussian mixture models.
15Information Geometry in Image Analysis (chaired by Yannick Berthoumieu, Geert Verdoolaege)Statistical Gaussian Model of Image Regions in Stochastic Watershed SegmentationJesús Angulo
Stochastic watershed is an image segmentation technique based on mathematical morphology which produces a probability density function of image contours. Estimated probabilities depend mainly on local distances between pixels. This paper introduces a variant of stochastic watershed where the probabilities of contours are computed from a gaussian model of image regions. In this framework, the basic ingredient is the distance between pairs of regions, hence a distance between normal distributions. Hence several alternatives of statistical distances for normal distributions are compared, namely Bhattacharyya distance, Hellinger metric distance and Wasserstein metric distance.
15Information Geometry in Image Analysis (chaired by Yannick Berthoumieu, Geert Verdoolaege)Quantization of hyperspectral image manifold using probabilistic distancesGianni Franchi, Jesús AnguloHyperspectral images, Information geometry, Mathematical morphology, Probabilistic distances, Quantization
A technique of spatial-spectral quantization of hyperspectral images is introduced. Thus a quantized hyperspectral image is just summarized by K spectra which represent the spatial and spectral structures of the image. The proposed technique is based on α-connected components on a region adjacency graph. The main ingredient is a dissimilarity metric. In order to choose the metric that best fit the hyperspectral data manifold, a comparison of different probabilistic dissimilarity measures is achieved.
16Optimal Transport and applications in Imagery/Statistics (chaired by Bertrand Maury, Jérémie Bigot)Non-convex relaxation of optimal transport for color transfer between imagesJulien Rabin, Nicolas PapadakisColor transfer, Optimal transport, Relaxation
Optimal transport (OT) is a major statistical tool to measure similarity between features or to match and average features. However, OT requires some relaxation and regularization to be robust to outliers. With relaxed methods, as one feature can be matched to several ones, important interpolations between different features arise. This is not an issue for comparison purposes, but it involves strong and unwanted smoothing for transfer applications. We thus introduce a new regularized method based on a non-convex formulation that minimizes transport dispersion by enforcing the one-to-one matching of features. The interest of the approach is demonstrated for color transfer purposes.
16Optimal Transport and applications in Imagery/Statistics (chaired by Bertrand Maury, Jérémie Bigot)Generalized Pareto Distributions, Image Statistics and Autofocusing in Automated MicroscopyReiner Lenz
We introduce the generalized Pareto distributions as a statistical model to describe thresholded edge-magnitude image filter results. Compared to the more commonWeibull or generalized extreme value distributions these distributions have at least two important advantages, the usage of the high threshold value assures that only the most important edge points enter the statistical analysis and the estimation is computationally more efficient since a much smaller number of data points have to be processed. The generalized Pareto distributions with a common threshold zero form a two-dimensional Riemann manifold with the metric given by the Fisher information matrix. We compute the Fisher matrix for shape parameters greater than -0.5 and show that the determinant of its inverse is a product of a polynomial in the shape parameter and the squared scale parameter. We apply this result by using the determinant as a sharpness function in an autofocus algorithm. We test the method on a large database of microscopy images with given ground truth focus results. We found that for a vast majority of the focus sequences the results are in the correct focal range. Cases where the algorithm fails are specimen with too few objects and sequences where contributions from different layers result in a multi-modal sharpness curve. Using the geometry of the manifold of generalized Pareto distributions more efficient autofocus algorithms can be constructed but these optimizations are not included here.
16Optimal Transport and applications in Imagery/Statistics (chaired by Bertrand Maury, Jérémie Bigot)Barycenter in Wasserstein space existence and consistencyJean-Michel Loubes, Thibaut Le GouicBarycenter, Geodesic spaces, Wasserstein space
We study barycenters in the Wasserstein space Pp(E) of a locally compact geodesic space (E, d). In this framework, we define the barycenter of a measure ℙ on Pp(E) as its Fréchet mean. The paper establishes its existence and states consistency with respect to ℙ. We thus extends previous results on ℝ d , with conditions on ℙ or on the sequence converging to ℙ for consistency.
16Optimal Transport and applications in Imagery/Statistics (chaired by Bertrand Maury, Jérémie Bigot)Multivariate L-moments based on transportsAlexis Decurninge
Univariate L-moments are expressed as projections of the quantile function onto an orthogonal basis of univariate polynomials. We present multivariate versions of L-moments expressed as collections of orthogonal projections of a multivariate quantile function on a basis of multivariate polynomials. We propose to consider quantile functions defined as transports from the uniform distribution on [0; 1] d onto the distribution of interest and present some properties of the subsequent L-moments. The properties of estimated L-moments are illustrated for heavy-tailed distributions.
17Probability Density Estimation (chaired by Jesús Angulo, S. Said)Probability density estimation on the hyperbolic space applied to radar processingEmmanuel Chevallier, Frédéric Barbaresco, Jesús Angulo
The two main techniques of probability density estimation on symmetric spaces are reviewed in the hyperbolic case. For computational reasons we chose to focus on the kernel density estimation and we provide the expression of Pelletier estimator on hyperbolic space. The method is applied to density estimation of reflection coefficients derived from radar observations.
17Probability Density Estimation (chaired by Jesús Angulo, S. Said)Histograms of images valued in the manifold of colours endowed with perceptual metricsEmmanuel Chevallier, Ivar Farup, Jesús AnguloColour images, Image histograms, Riemannian metrics
We address here the problem of perceptual colour histograms. The Riemannian structure of perceptual distances is measured through standards sets of ellipses, such as Macadam ellipses. We propose an approach based on local Euclidean approximations that enables to take into account the Riemannian structure of perceptual distances, without introducing computational complexity during the construction of the histogram.
17Probability Density Estimation (chaired by Jesús Angulo, S. Said)Entropy minimizing curves with application to automated flight path designFlorence Nicol, Stephane Puechmorel
Air traffic management (ATM) aims at providing companies with a safe and ideally optimal aircraft trajectory planning. Air traffic controllers act on flight paths in such a way that no pair of aircraft come closer than the regulatory separation norm. With the increase of traffic, it is expected that the system will reach its limits in a near future: a paradigm change in ATM is planned with the introduction of trajectory based operations. This paper investigate a mean of producing realistic air routes from the output of an automated trajectory design tool. For that purpose, an entropy associated with a system of curves is defined and a mean of iteratively minimizing it is presented. The network produced is suitable for use in a semi-automated ATM system with human in the loop.
17Probability Density Estimation (chaired by Jesús Angulo, S. Said)Kernel Density Estimation on Symmetric SpacesDena Asta
We introduce a novel kernel density estimator for a large class of symmetric spaces and prove a minimax rate of convergence as fast as the minimax rate on Euclidean space. We prove a minimax rate of convergence proven without any compactness assumptions on the space or Hölder-class assumptions on the densities. A main tool used in proving the convergence rate is the Helgason-Fourier transform, a generalization of the Fourier transform for semisimple Lie groups modulo maximal compact subgroups. This paper obtains a simplified formula in the special case when the symmetric space is the 2-dimensional hyperboloid.
21Keynote speach Tudor Ratiu (chaired by Xavier Pennec)Symetry methods in geometrics mechanicsTudor Ratiu
The goal of these lectures is to show the influence of symmetry in various aspects of theoretical mechanics. Canonical actions of Lie groups on Poisson manifolds often give rise to conservation laws, encoded in modern language by the concept of momentum maps. Reduction methods lead to a deeper understanding of the dynamics of mechanical systems. Basic results in singular Hamiltonian reduction will be presented. The Lagrangian version of reduction and its associated variational principles will also be discussed. The understanding of symmetric bifurcation phenomena in for Hamiltonian systems are based on these reduction techniques. Time permitting, discrete versions of these geometric methods will also be discussed in the context of examples from elasticity. 
22Dimension reduction on Riemannian manifolds (chaired by Xavier Pennec, Alain Trouvé)Evolution Equations with Anisotropic Distributions and Diffusion PCAStefan Sommer
This paper presents derivations of evolution equations for the family of paths that in the Diffusion PCA framework are used for approximating data likelihood. The paths that are formally interpreted as most probable paths generalize geodesics in extremizing an energy functional on the space of differentiable curves on a manifold with connection. We discuss how the paths arise as projections of geodesics for a (non bracket-generating) sub-Riemannian metric on the frame bundle. Evolution equations in coordinates for both metric and cometric formulations of the sub-Riemannian geometry are derived. We furthermore show how rank-deficient metrics can be mixed with an underlying Riemannian metric, and we use the construction to show how the evolution equations can be implemented on finite dimensional LDDMM landmark manifolds.
22Dimension reduction on Riemannian manifolds (chaired by Xavier Pennec, Alain Trouvé)Barycentric Subspaces and Affine Spans in ManifoldsXavier Pennec
This paper addresses the generalization of Principal Component Analysis (PCA) to Riemannian manifolds. Current methods like Principal Geodesic Analysis (PGA) and Geodesic PCA (GPCA) minimize the distance to a “Geodesic subspace”. This allows to build sequences of nested subspaces which are consistent with a forward component analysis approach. However, these methods cannot be adapted to a backward analysis and they are not symmetric in the parametrization of the subspaces. We propose in this paper a new and more general type of family of subspaces in manifolds: barycentric subspaces are implicitly defined as the locus of points which are weighted means of k + 1 reference points. Depending on the generalization of the mean that we use, we obtain the Fréchet/Karcher barycentric subspaces (FBS/KBS) or the affine span (with exponential barycenter). This definition restores the full symmetry between all parameters of the subspaces, contrarily to the geodesic subspaces which intrinsically privilege one point. We show that this definition defines locally a submanifold of dimension k and that it generalizes in some sense geodesic subspaces. Like PGA, barycentric subspaces allow the construction of a forward nested sequence of subspaces which contains the Fréchet mean. However, the definition also allows the construction of backward nested sequence which may not contain the mean. As this definition relies on points and do not explicitly refer to tangent vectors, it can be extended to non Riemannian geodesic spaces. For instance, principal subspaces may naturally span over several strata in stratified spaces, which is not the case with more classical generalizations of PCA.
22Dimension reduction on Riemannian manifolds (chaired by Xavier Pennec, Alain Trouvé)Dimension Reduction on Polyspheres with Application to Skeletal RepresentationsBenjamin Eltzner, Stephan Huckemann, Sungkyu Jung
We present a novel method that adaptively deforms a polysphere (a product of spheres) into a single high dimensional sphere which then allows for principal nested spheres (PNS) analysis. Applying our method to skeletal representations of simulated bodies as well as of data from real human hippocampi yields promising results in view of dimension reduction. Specifically in comparison to composite PNS (CPNS), our method of principal nested deformed spheres (PNDS) captures essential modes of variation by lower dimensional representations.
22Dimension reduction on Riemannian manifolds (chaired by Xavier Pennec, Alain Trouvé)Affine-invariant Riemannian Distance Between Infinite-dimensional Covariance OperatorsMinh Ha Quang
This paper studies the affine-invariant Riemannian distance on the Riemann-Hilbert manifold of positive definite operators on a separable Hilbert space. This is the generalization of the Riemannian manifold of symmetric, positive definite matrices to the infinite-dimensional setting. In particular, in the case of covariance operators in a Reproducing Kernel Hilbert Space (RKHS), we provide a closed form solution, expressed via the corresponding Gram matrices.
22Dimension reduction on Riemannian manifolds (chaired by Xavier Pennec, Alain Trouvé)A sub-Riemannian modular approach for diffeomorphic deformationsAlain Trouvé, Barbara Gris, Stanley Durrleman
We develop a generic framework to build large deformations from a combination of base modules. These modules constitute a dynamical dictionary to describe transformations. The method, built on a coherent sub-Riemannian framework, defines a metric on modular deformations and characterises optimal deformations as geodesics for this metric. We will present a generic way to build local affine transformations as deformation modules, and display examples.
23Optimization on Manifold (chaired by Pierre-Antoine Absil, Rodolphe Sepulchre)Riemannian trust regions with finite-difference Hessian approximations are globally convergentNicolas BoumalConvergence, Manopt, Optimization on manifolds, RTR-FD
The Riemannian trust-region algorithm (RTR) is designed to optimize differentiable cost functions on Riemannian manifolds. It proceeds by iteratively optimizing local models of the cost function. When these models are exact up to second order, RTR boasts a quadratic convergence rate to critical points. In practice, building such models requires computing the Riemannian Hessian, which may be challenging. A simple idea to alleviate this difficulty is to approximate the Hessian using finite differences of the gradient. Unfortunately, this is a nonlinear approximation, which breaks the known convergence results for RTR. We propose RTR-FD: a modification of RTR which retains global convergence when the Hessian is approximated using finite differences. Importantly, RTR-FD reduces gracefully to RTR if a linear approximation is used. This algorithm is available in the Manopt toolbox.
23Optimization on Manifold (chaired by Pierre-Antoine Absil, Rodolphe Sepulchre)Block-Jacobi methods with Newton-steps and non-unitary joint matrix diagonalizationHao Shen, Martin KleinsteuberJacobi algorithms, Local convergence properties, Manifold optimization, Signal separation
In this work, we consider block-Jacobi methods with Newton steps in each subspace search and prove their local quadratic convergence to a local minimum with non-degenerate Hessian under some orthogonality assumptions on the search directions. Moreover, such a method is exemplified for non-unitary joint matrix diagonalization, where we present a block-Jacobi-type method on the oblique manifold with guaranteed local quadratic convergence.
23Optimization on Manifold (chaired by Pierre-Antoine Absil, Rodolphe Sepulchre)Weakly Correlated Sparse Components with Nearly Orthonormal LoadingsMatthieu Genicot, Nickolay Trendafilov, Wen HuangDantzig selector, LASSO, Optimization on matrix manifolds, Orthonormal and oblique component loadings matrices
There is already a great number of highly efficient methods producing components with sparse loadings which significantly facilitates the interpretation of principal component analysis (PCA). However, they produce either only orthonormal loadings, or only uncorrelated components, or, most frequently, neither of them. To overcome this weakness, we introduce a new approach to define sparse PCA similar to the Dantzig selector idea already employed for regression problems. In contrast to the existing methods, the new approach makes it possible to achieve simultaneously nearly uncorrelated sparse components with nearly orthonormal loadings. The performance of the new method is illustrated on real data sets. It is demonstrated that the new method outperforms one of the most popular available methods for sparse PCA in terms of preservation of principal components properties.
23Optimization on Manifold (chaired by Pierre-Antoine Absil, Rodolphe Sepulchre)Fitting Smooth Paths on Riemannian Manifolds - Endometrial SurfaceAntoine Arnould, Chafik Samir, Michel Canis, Pierre-Antoine Absil, Pierre-Yves GousenbourgerBézier functions, Endometrial surface reconstruction, MRI-based navigation, Optimization on manifolds, Path fitting on Riemannian manifolds
We present a new method to fit smooth paths to a given set of points on Riemannian manifolds using C1 piecewise-Bézier functions. A property of the method is that, when the manifold reduces to a Euclidean space, the control points minimize the mean square acceleration of the path. As an application, we focus on data observations that evolve on certain nonlinear manifolds of importance in medical imaging: the shape manifold for endometrial surface reconstruction; the special orthogonal group SO(3) and the special Euclidean group SE(3) for preoperative MRI-based navigation. Results on real data show that our method succeeds in meeting the clinical goal: combining different modalities to improve the localization of the endometrial lesions.
23Optimization on Manifold (chaired by Pierre-Antoine Absil, Rodolphe Sepulchre)PDE Constrained Shape Optimization as Optimization on Shape ManifoldsKathrin Welker, Martin Siebenborn, Volker SchulzLimited memory BFGS, Newton method, Quasi–Newton method, Riemannian manifold, Shape optimization
The novel Riemannian view on shape optimization introduced in [14] is extended to a Lagrange–Newton as well as a quasi–Newton approach for PDE constrained shape optimization problems.
24Shape Space & Diffeomorphic mappings (chaired by Stanley Durrleman, Stéphanie Allassonnière)Spherical parameterization for genus zero surfaces using Laplace-Beltrami eigenfunctionsGuillaume Auzias, Julien LefèvreLaplace-Beltrami Operator, Nodal domains, Riemannian manifold, Surface parameterization
In this work, we propose a fast and simple approach to obtain a spherical parameterization of a certain class of closed surfaces without holes. Our approach relies on empirical findings that can be mathematically investigated, to a certain extent, by using Laplace-Beltrami Operator and associated geometrical tools. The mapping proposed here is defined by considering only the three first non-trivial eigenfunctions of the Laplace-Beltrami Operator. Our approach requires a topological condition on those eigenfunctions, whose nodal domains must be 2. We show the efficiency of the approach through numerical experiments performed on cortical surface meshes.
24Shape Space & Diffeomorphic mappings (chaired by Stanley Durrleman, Stéphanie Allassonnière)Biased estimators on quotient spacesNina Miolane, Xavier Pennec
Usual statistics are defined, studied and implemented on Euclidean spaces. But what about statistics on other mathematical spaces, like manifolds with additional properties: Lie groups, Quotient spaces, Stratified spaces etc? How can we describe the interaction between statistics and geometry? The structure of Quotient space in particular is widely used to model data, for example every time one deals with shape data. These can be shapes of constellations in Astronomy, shapes of human organs in Computational Anatomy, shapes of skulls in Palaeontology, etc. Given this broad field of applications, statistics on shapes -and more generally on observations belonging to quotient spaces- have been studied since the 1980’s. However, most theories model the variability in the shapes but do not take into account the noise on the observations themselves. In this paper, we show that statistics on quotient spaces are biased and even inconsistent when one takes into account the noise. In particular, some algorithms of template estimation in Computational Anatomy are biased and inconsistent. Our development thus gives a first theoretical geometric explanation of an experimentally observed phenomenon. A biased estimator is not necessarily a problem. In statistics, it is a general rule of thumb that a bias can be neglected for example when it represents less than 0.25 of the variance of the estimator. We can also think about neglecting the bias when it is low compared to the signal we estimate. In view of the applications, we thus characterize geometrically the situations when the bias can be neglected with respect to the situations when it must be corrected.
24Shape Space & Diffeomorphic mappings (chaired by Stanley Durrleman, Stéphanie Allassonnière)Uniqueness of the Fisher-Rao Metric on the Space of Smooth DensitiesMartin Bauer, Martins Bruveris, Peter Michor
We review the manifold projection method for stochastic nonlinear filtering in a more general setting than in our previous paper in Geometric Science of Information 2013. We still use a Hilbert space structure on a space of probability densities to project the infinite dimensional stochastic partial differential equation for the optimal filter onto a finite dimensional exponential or mixture family, respectively, with two different metrics, the Hellinger distance and the L2 direct metric. This reduces the problem to finite dimensional stochastic differential equations. In this paper we summarize a previous equivalence result between Assumed Density Filters (ADF) and Hellinger/Exponential projection filters, and introduce a new equivalence between Galerkin method based filters and Direct metric/Mixture projection filters. This result allows us to give a rigorous geometric interpretation to ADF and Galerkin filters. We also discuss the different finite-dimensional filters obtained when projecting the stochastic partial differential equation for either the normalized (Kushner-Stratonovich) or a specific unnormalized (Zakai) density of the optimal filter.
24Shape Space & Diffeomorphic mappings (chaired by Stanley Durrleman, Stéphanie Allassonnière)Reparameterization invariant metric on the space of curvesAlice Le Brigant, Frédéric Barbaresco, Marc Arnaudon
This paper focuses on the study of open curves in a manifold M, and its aim is to define a reparameterization invariant distance on the space of such paths. We use the square root velocity function (SRVF) introduced by Srivastava et al. in [11] to define a reparameterization invariant metric on the space of immersions =Imm([0,1],M) by pullback of a metric on the tangent bundle T derived from the Sasaki metric. We observe that such a natural choice of Riemannian metric on T induces a first-order Sobolev metric on with an extra term involving the origins, and leads to a distance which takes into account the distance between the origins and the distance between the image curves by the SRVF parallel transported to a same vector space, with an added curvature term. This provides a generalized theoretical SRV framework for curves lying in a general manifold M.
24Shape Space & Diffeomorphic mappings (chaired by Stanley Durrleman, Stéphanie Allassonnière)Invariant geometric structures on statistical modelsHong Van Le, Jürgen Jost, Lorenz Schwachhöfer, Nihat Ay
We review the notion of parametrized measure models and tensor fields on them, which encompasses all statistical models considered by Chentsov [6], Amari [3] and Pistone-Sempi [10]. We give a complete description of n-tensor fields that are invariant under sufficient statistics. In the cases n = 2 and n = 3, the only such tensors are the Fisher metric and the Amari-Chentsov tensor. While this has been shown by Chentsov [7] and Campbell [5] in the case of finite measure spaces, our approach allows to generalize these results to the cases of infinite sample spaces and arbitrary n. Furthermore, we give a generalisation of the monotonicity theorem and discuss its consequences.
24Shape Space & Diffeomorphic mappings (chaired by Stanley Durrleman, Stéphanie Allassonnière)The abstract setting for shape deformation analysis and LDDMM methodsSylvain Arguillere
This paper aims to define a unified setting for shape registration and LDDMM methods for shape analysis. This setting turns out to be sub-Riemannian, and not Riemannian. An abstract definition of a space of shapes in ℝ d is given, and the geodesic flow associated to any reproducing kernel Hilbert space of sufficiently regular vector fields is showed to exist for all time.
25Lie Groups: Novel Statistical and Computational Frontiers (chaired by Jérémie Jakubowicz, Bijan Afsari)An intrinsic Cramér-Rao bound on Lie groupsAxel Barrau, Silvère Bonnabel
In his 2005 paper, S.T. Smith proposed an intrinsic Cramér- Rao bound on the variance of estimators of a parameter defined on a Riemannian manifold. In the present technical note, we consider the special case where the parameter lives in a Lie group. In this case, by choosing, e.g., the right invariant metric, parallel transport becomes very simple, which allows a more straightforward and natural derivation of the bound in terms of Lie bracket, albeit for a slightly different definition of the estimation error. For bi-invariant metrics, the Lie group exponential map we use to define the estimation error, and the Riemannian exponential map used by S.T. Smith coincide, and we prove in this case that both results are identical indeed.
25Lie Groups: Novel Statistical and Computational Frontiers (chaired by Jérémie Jakubowicz, Bijan Afsari)Image processing in the semidiscrete group of rototranslationsDario Prandi, Jean-Paul Gauthier, Ugo Boscain
It is well-known, since [12], that cells in the primary visual cortex V1 do much more than merely signaling position in the visual field: most cortical cells signal the local orientation of a contrast edge or bar . they are tuned to a particular local orientation. This orientation tuning has been given a mathematical interpretation in a sub-Riemannian model by Petitot, Citti, and Sarti [6,14]. According to this model, the primary visual cortex V1 lifts grey-scale images, given as functions f : ℝ2 → [0, 1], to functions Lf defined on the projectivized tangent bundle of the plane PTℝ2 = ℝ2 ×ℙ1. Recently, in [1], the authors presented a promising semidiscrete variant of this model where the Euclidean group of rototranslations SE(2), which is the double covering of PTℝ2, is replaced by SE(2,N), the group of translations and discrete rotations. In particular, in [15], an implementation of this model allowed for state-of-the-art image inpaintings. In this work, we review the inpainting results and introduce an application of the semidiscrete model to image recognition. We remark that both these applications deeply exploit the Moore structure of SE(2,N) that guarantees that its unitary representations behaves similarly to those of a compact group. This allows for nice properties of the Fourier transform on SE(2,N) exploiting which one obtains numerical advantages.
25Lie Groups: Novel Statistical and Computational Frontiers (chaired by Jérémie Jakubowicz, Bijan Afsari)Universal, non-asymptotic confidence sets for circular meansFlorian Kelma, Johannes Wieditz, Thomas Hotz
Based on Hoeffding’s mass concentration inequalities, nonasymptotic confidence sets for circular means are constructed which are universal in the sense that they require no distributional assumptions. These are then compared with asymptotic confidence sets in simulations and for a real data set.
25Lie Groups: Novel Statistical and Computational Frontiers (chaired by Jérémie Jakubowicz, Bijan Afsari)Deblurring and Recovering Conformational States in 3D Single Particle ElectronBijan Afsari, Gregory S. Chirikjian
In this paper we study two forms of blurring effects that may appear in the reconstruction of 3D Electron Microscopy (EM), specifically in single particle reconstruction from random orientations of large multi-unit biomolecular complexes. We model the blurring effects as being due to independent contributions from: (1) variations in the conformation of the biomolecular complex; and (2) errors accumulated in the reconstruction process. Under the assumption that these effects can be separated and treated independently, we show that the overall blurring effect can be expressed as a special form of a convolution operation of the 3D density with a kernel defined on SE(3), the Lie group of rigid body motions in 3D. We call this form of convolution mixed spatial-motional convolution.We discuss the ill-conditioned nature of the deconvolution needed to deblur the reconstructed 3D density in terms of parameters associated with the unknown probability in SE(3). We provide an algorithm for recovering the conformational information of large multi-unit biomolecular complexes (essentially deblurring) under certain biologically plausible prior structural knowledge about the subunits of the complex in the case the blurring kernel has a special form.
25Lie Groups: Novel Statistical and Computational Frontiers (chaired by Jérémie Jakubowicz, Bijan Afsari)Nonlinear operators on graphs via stacksJesús Angulo, Santiago Velasco-Forero
We consider a framework for nonlinear operators on functions evaluated on graphs via stacks of level sets. We investigate a family of transformations on functions evaluated on graph which includes adaptive flat and non-flat erosions and dilations in the sense of mathematical morphology. Additionally, the connection to mean motion curvature on graphs is noted. Proposed operators are illustrated in the cases of functions on graphs, textured meshes and graphs of images.
26Lie Groups and Geometric Mechanics/Thermodynamics (chaired by Frédéric Barbaresco, Géry de Saxcé)Poincaré's equations for Cosserat shells - application to locomotion of cephalopodsFederico Renda, Frederic Boyer
In 1901 Henri Poincaré proposed a new set of equations for mechanics. These equations are a generalization of Lagrange equations to a system whose configuration space is a Lie group, which is not necessarily commutative. Since then, this result has been extensively refined by the Lagrangian reduction theory. In this article, we show the relations between these equations and continuous Cosserat media, i.e. media for which the conventional model of point particle is replaced by a rigid body of small volume named microstructure. In particular, we will see that the usual shell balance equations of nonlinear structural dynamics can be easily derived from the Poincaré’s result. This framework is illustrated through the simulation of a simplified model of cephalopod swimming.
26Lie Groups and Geometric Mechanics/Thermodynamics (chaired by Frédéric Barbaresco, Géry de Saxcé)Entropy and structure of the thermodynamical systemsGéry de Saxcé
With respect to the concept of affine tensor, we analyse in this work the underlying geometric structure of the theories of Lie group statistical mechanics and relativistic thermodynamics of continua, formulated by Souriau independently one of each other. We reveal the link between these ones in the classical Galilean context. These geometric structures of the thermodynamics are rich and we think they might be source of inspiration for the geometric theory of information based on the concept of entropy.
26Lie Groups and Geometric Mechanics/Thermodynamics (chaired by Frédéric Barbaresco, Géry de Saxcé)Symplectic Structure of Information Geometry: Fisher Metric and Euler-Poincaré Equation of Souriau Lie Group ThermodynamicsFrédéric BarbarescoCartan-Poincaré Integral, Euler-Poincaré Equation, Fisher Metric, Gauge Theory, Geometric Mechanics, Gibbs Equilibrium, Information geometry, Invariant Lie Group, Maximum Entropy, Momentum Map, Symplectic Geometry, Thermodynamics
We introduce the Symplectic Structure of Information Geometry based on Souriau’s Lie Group Thermodynamics model, with a covariant definition of Gibbs equilibrium via invariances through co-adjoint action of a group on its momentum space, defining physical observables like energy, heat, and momentum as pure geometrical objects. Using Geometric (Planck) Temperature of Souriau model and Symplectic cocycle notion, the Fisher metric is identified as a Souriau Geometric Heat Capacity. In the framework of Lie Group Thermodynamics, an Euler-Poincaré equation is elaborated with respect to thermodynamic variables, and a new variational principal for thermodynamics is built through an invariant Poincaré-Cartan-Souriau integral. Finally, we conclude on Balian Gauge theory of Thermodynamics compatible with Souriau’s Model.
26Lie Groups and Geometric Mechanics/Thermodynamics (chaired by Frédéric Barbaresco, Géry de Saxcé)Pontryagin calculus in Riemannian geometryDanielle Fortune, Francois Dubois, Juan Antonio Rojas QuinteroEuler-Lagrange equation, Riemann curvature tensor, Robotics
In this contribution, we study systems with a finite number of degrees of freedom as in robotics. A key idea is to consider the mass tensor associated to the kinetic energy as a metric in a Riemannian configuration space. We apply Pontryagin’s framework to derive an optimal evolution of the control forces and torques applied to the mechanical system. This equation under covariant form uses explicitly the Riemann curvature tensor.
26Lie Groups and Geometric Mechanics/Thermodynamics (chaired by Frédéric Barbaresco, Géry de Saxcé)Rolling Symmetric SpacesFátima Leite, Krzysztof Krakowski, Luís MachadoGraßmann manifold, Lie algebra, Rolling Isometry, Symmetric spaces
Riemannian symmetric spaces play an important role in many areas that are interrelated to information geometry. For instance, in image processing one of the most elementary tasks is image interpolation. Since a set of images may be represented by a point in the Graßmann manifold, image interpolation can be formulated as an interpolation problem on that symmetric space. It turns out that rolling motions, subject to nonholonomic constraints of no-slip and no-twist, provide efficient algorithms to generate interpolating curves on certain Riemannian manifolds, in particular on symmetric spaces. The main goal of this paper is to study rolling motions on symmetric spaces. It is shown that the natural decomposition of the Lie algebra associated to a symmetric space provides the structure of the kinematic equations that describe the rolling motion of that space upon its affine tangent space at a point. This generalizes what can be observed in all the particular cases that are known to the authors. Some of these cases illustrate the general results.
26Lie Groups and Geometric Mechanics/Thermodynamics (chaired by Frédéric Barbaresco, Géry de Saxcé)Enlargement, geodesics, and collectivesEric Justh, P. S. KrishnaprasadCollective behavior, Lie-poisson reduction, Nonholonomic integrator, Subriemannian geodesics, Subriemannian geometry
We investigate optimal control of systems of particles on matrix Lie groups coupled through graphs of interaction, and characterize the limit of strong coupling. Following Brockett, we use an enlargement approach to obtain a convenient form of the optimal controls. In the setting of drift-free particle dynamics, the coupling terms in the cost functionals lead to a novel class of problems in subriemannian geometry of product Lie groups.
27Divergence Geometry (chaired by Michel Broniatowski, Imre Csiszár)Generalized EM algorithms for minimum divergence estimationDiaa Al Mohamad, Michel Broniatowski
Minimum divergence estimators are derived through the dual form of the divergence in parametric models. These estimators generalize the classical maximum likelihood ones. Models with unobserved data, as mixture models, can be estimated with EM algorithms, which are proved to converge to stationary points of the likelihood function under general assumptions. This paper presents an extension of the EM algorithm based on minimization of the dual approximation of the divergence between the empirical measure and the model using a proximaltype algorithm. The algorithm converges to the stationary points of the empirical criterion under general conditions pertaining to the divergence and the model. Robustness properties of this algorithm are also presented. We provide another proof of convergence of the EM algorithm in a two-component gaussian mixture. Simulations on Gaussian andWeibull mixtures are performed to compare the results with the MLE.
27Divergence Geometry (chaired by Michel Broniatowski, Imre Csiszár)Extension of information geometry to non-statistical systems some examplesBen Anthonis, Jan Naudts, Michel Broniatowski
Our goal is to extend information geometry to situations where statistical modeling is not obvious. The setting is that of modeling experimental data. Quite often the data are not of a statistical nature. Sometimes also the model is not a statistical manifold. An example of the former is the description of the Bose gas in the grand canonical ensemble. An example of the latter is the modeling of quantum systems with density matrices. Conditional expectations in the quantum context are reviewed. The border problem is discussed: through conditioning the model point shifts to the border of the differentiable manifold.
27Divergence Geometry (chaired by Michel Broniatowski, Imre Csiszár)An Information Geometry Problem in Mathematical FinanceImre Csiszár, Michel Broniatowski, Thomas BreuerAlmost worst case densities, Bregman distance, Convex integral functional, f-divergence I-divergence, Payoff function, Pythagorean identity, Risk measure, Worst case density
Familiar approaches to risk and preferences involve minimizing the expectation EIP(X) of a payoff function X over a family Γ of plausible risk factor distributions IP. We consider Γ determined by a bound on a convex integral functional of the density of IP, thus Γ may be an I-divergence (relative entropy) ball or some other f-divergence ball or Bregman distance ball around a default distribution IPo. Using a Pythagorean identity we show that whether or not a worst case distribution exists (minimizing EIP(X) subject to IP∈Γ), the almost worst case distributions cluster around an explicitly specified, perhaps incomplete distribution. When Γ is an f-divergence ball, a worst case distribution either exists for any radius, or it does/does not exist for radius less/larger than a critical value. It remains open how far the latter result extends beyond f-divergence balls.
27Divergence Geometry (chaired by Michel Broniatowski, Imre Csiszár)Multivariate divergences with application in multisample density ratio modelsAmor Keziou
We introduce what we will call multivariate divergences between K, K ≥ 1, signed finite measures (Q1, . . . , Q K ) and a given reference probability measure P on a σ-field (X,B), extending the well known divergences between two measures, a signed finite measure Q1 and a given probability distribution P. We investigate the Fenchel duality theory for the introduced multivariate divergences viewed as convex functionals on well chosen topological vector spaces of signed finite measures. We obtain new dual representations of these criteria, which we will use to define new family of estimates and test statistics with multiple samples under multiple semiparametric density ratio models. This family contains the estimate and test statistic obtained through empirical likelihood. Moreover, the present approach allows obtaining the asymptotic properties of the estimates and test statistics both under the model and under misspecification. This leads to accurate approximations of the power function for any used criterion, including the empirical likelihood one, which is of its own interest. Moreover, the proposed multivariate divergences can be used, in the context of multiple samples in density ratio models, to define new criteria for model selection and multi-group classification.
27Divergence Geometry (chaired by Michel Broniatowski, Imre Csiszár)Generalized Mutual-Information based independence testsAmor Keziou, Philippe Regnault
We derive independence tests by means of dependence measures thresholding in a semiparametric context. Precisely, estimates of mutual information associated to ϕ-divergences are derived through the dual representations of ϕ-divergences. The asymptotic properties of the estimates are established, including consistency, asymptotic distribution and large deviations principle. The related tests of independence are compared through their relative asymptotic Bahadur efficiency and numerical simulations.
28Invited speaker Charles-Michel Marle (chaired by Frédéric Barbaresco)Actions of Lie groups and Lie algebras on symplectic and Poisson manifolds. Application to Lagrangian and Hamiltonian systemsCharles-Michel Marle
I will present some tools in Symplectic and Poisson Geometry in view of their applications in Geometric Mechanics and Mathematical Physics. Lie group and Lie algebra actions on symplectic and Poisson manifolds, momentum maps and their equivariance properties, first integrals associated to symmetries of Hamiltonian systems will be discussed. Reduction methods taking advantage of symmetries will be discussed.