arXiv subjects

Frank Nielsen

Publications and source records attributed to Frank Nielsen.

137 records · Page 8Linked to original sources

arXiv2011

On Rényi and Tsallis entropies and divergences for exponential families

Many common probability distributions in statistics like the Gaussian, multinomial, Beta or Gamma distributions can be studied under the unified framework of exponential families. In this paper, we prove that both Rényi and Tsallis divergences of distributions belonging to the same exponential family admit a generic closed form expression. Furthermore, we show that Rényi and Tsallis entropies can also be calculated in closed-form for sub-families including the Gaussian or exponential distributions, among others.

cs.IT↗

arXiv2011

Statistical exponential families: A digest with flash cards

This document describes concisely the ubiquitous class of exponential family distributions met in statistics. The first part recalls definitions and summarizes main properties and duality with Bregman divergences (all proofs are skipped). The second part lists decompositions and related formula of common exponential family distributions. We recall the Fisher-Rao-Riemannian geometries and the dual affine connection information geometries of statistical manifolds. It is intended to maintain and update this document and catalog by adding new distribution items.

cs.LG↗

arXiv2011

Chernoff information of exponential families

Chernoff information upper bounds the probability of error of the optimal Bayesian decision rule for $2$-class classification problems. However, it turns out that in practice the Chernoff bound is hard to calculate or even approximate. In statistics, many usual distributions, such as Gaussians, Poissons or frequency histograms called multinomials, can be handled in the unified framework of exponential families. In this note, we prove that the Chernoff information for members of the same exponential family can be either derived analytically in closed form, or efficiently approximated using a simple geodesic bisection optimization technique based on an exact geometric characterization of the "Chernoff point" on the underlying statistical manifold.

cs.IT↗

arXiv2010

Video Stippling

In this paper, we consider rendering color videos using a non-photo-realistic art form technique commonly called stippling. Stippling is the art of rendering images using point sets, possibly with various attributes like sizes, elementary shapes, and colors. Producing nice stippling is attractive not only for the sake of image depiction but also because it yields a compact vectorial format for storing the semantic information of media. Moreover, stippling is by construction easily tunable to various device resolutions without suffering from bitmap sampling artifacts when resizing. The underlying core technique for stippling images is to compute a centroidal Voronoi tessellation on a well-designed underlying density. This density relates to the image content, and is used to compute a weighted Voronoi diagram. By considering videos as image sequences and initializing properly the stippling of one image by the result of its predecessor, one avoids undesirable point flickering artifacts and can produce stippled videos that nevertheless still exhibit noticeable artifacts. To overcome this, our method improves over the naive scheme by considering dynamic point creation and deletion according to the current scene semantic complexity, and show how to effectively vectorize video while adjusting for both color and contrast characteristics. Furthermore, we explain how to produce high quality stippled ``videos'' (eg., fully dynamic spatio-temporal point sets) for media containing various fading effects, like quick motions of objects or progressive shot changes. We report on practical performances of our implementation, and present several stippled video results rendered on-the-fly using our viewer that allows both spatio-temporal dynamic rescaling (eg., upscale vectorially frame rate).

cs.GR↗

arXiv2010

Boosting k-NN for categorization of natural scenes

The k-nearest neighbors (k-NN) classification rule has proven extremely successful in countless many computer vision applications. For example, image categorization often relies on uniform voting among the nearest prototypes in the space of descriptors. In spite of its good properties, the classic k-NN rule suffers from high variance when dealing with sparse prototype datasets in high dimensions. A few techniques have been proposed to improve k-NN classification, which rely on either deforming the nearest neighborhood relationship or modifying the input space. In this paper, we propose a novel boosting algorithm, called UNN (Universal Nearest Neighbors), which induces leveraged k-NN, thus generalizing the classic k-NN rule. We redefine the voting rule as a strong classifier that linearly combines predictions from the k closest prototypes. Weak classifiers are learned by UNN so as to minimize a surrogate risk. A major feature of UNN is the ability to learn which prototypes are the most relevant for a given class, thus allowing one for effective data reduction. Experimental results on the synthetic two-class dataset of Ripley show that such a filtering strategy is able to reject "noisy" prototypes. We carried out image categorization experiments on a database containing eight classes of natural scenes. We show that our method outperforms significantly the classic k-NN classification, while enabling significant reduction of the computational cost by means of data filtering.

cs.CV↗

arXiv2009

Hyperbolic Voronoi diagrams made easy

We present a simple framework to compute hyperbolic Voronoi diagrams of finite point sets as affine diagrams. We prove that bisectors in Klein's non-conformal disk model are hyperplanes that can be interpreted as power bisectors of Euclidean balls. Therefore our method simply consists in computing an equivalent clipped power diagram followed by a mapping transformation depending on the selected representation of the hyperbolic space (e.g., Poincaré conformal disk or upper-plane representations). We discuss on extensions of this approach to weighted and $k$-order diagrams, and describe their dual triangulations. Finally, we consider two useful primitives on the hyperbolic Voronoi diagrams for designing tailored user interfaces of an image catalog browsing application in the hyperbolic disk: (1) finding nearest neighbors, and (2) computing smallest enclosing balls.

cs.CG↗

arXiv2009

Information geometries and Microeconomic Theories

More than thirty years ago, Charnes, Cooper and Schinnar (1976) established an enlightening contact between economic production functions (EPFs) -- a cornerstone of neoclassical economics -- and information theory, showing how a generalization of the Cobb-Douglas production function encodes homogeneous functions. As expected by Charnes \textit{et al.}, the contact turns out to be much broader: we show how information geometry as pioneered by Amari and others underpins static and dynamic descriptions of microeconomic cornerstones. We show that the most popular EPFs are fundamentally grounded in a very weak axiomatization of economic transition costs between inputs. The strength of this characterization is surprising, as it geometrically bonds altogether a wealth of collateral economic notions -- advocating for applications in various economic fields --: among all, it characterizes (i) Marshallian and Hicksian demands and their geometric duality, (ii) Slutsky-type properties for the transformation paths, (iii) Roy-type properties for their elementary variations.

q-fin.GN↗

arXiv2008

Soft Uncoupling of Markov Chains for Permeable Language Distinction: A New Algorithm

Without prior knowledge, distinguishing different languages may be a hard task, especially when their borders are permeable. We develop an extension of spectral clustering -- a powerful unsupervised classification toolbox -- that is shown to resolve accurately the task of soft language distinction. At the heart of our approach, we replace the usual hard membership assignment of spectral clustering by a soft, probabilistic assignment, which also presents the advantage to bypass a well-known complexity bottleneck of the method. Furthermore, our approach relies on a novel, convenient construction of a Markov chain out of a corpus. Extensive experiments with a readily available system clearly display the potential of the method, which brings a visually appealing soft distinction of languages that may define altogether a whole corpus.

cs.CL↗

arXiv2008

Staring at Economic Aggregators through Information Lenses

It is hard to exaggerate the role of economic aggregators -- functions that summarize numerous and / or heterogeneous data -- in economic models since the early XX$^{th}$ century. In many cases, as witnessed by the pioneering works of Cobb and Douglas, these functions were information quantities tailored to economic theories, i.e. they were built to fit economic phenomena. In this paper, we look at these functions from the complementary side: information. We use a recent toolbox built on top of a vast class of distortions coined by Bregman, whose application field rivals metrics' in various subfields of mathematics. This toolbox makes it possible to find the quality of an aggregator (for consumptions, prices, labor, capital, wages, etc.), from the standpoint of the information it carries. We prove a rather striking result. From the informational standpoint, well-known economic aggregators do belong to the \textit{optimal} set. As common economic assumptions enter the analysis, this large set shrinks, and it essentially ends up \textit{exactly fitting} either CES, or Cobb-Douglas, or both. To summarize, in the relevant economic contexts, one could not have crafted better some aggregator from the information standpoint. We also discuss global economic behaviors of optimal information aggregators in general, and present a brief panorama of the links between economic and information aggregators. Keywords: Economic Aggregators, CES, Cobb-Douglas, Bregman divergences

cs.IT↗

arXiv2007

On the Centroids of Symmetrized Bregman Divergences

In this paper, we generalize the notions of centroids and barycenters to the broad class of information-theoretic distortion measures called Bregman divergences. Bregman divergences are versatile, and unify quadratic geometric distances with various statistical entropic measures. Because Bregman divergences are typically asymmetric, we consider both the left-sided and right-sided centroids and the symmetrized centroids, and prove that all three are unique. We give closed-form solutions for the sided centroids that are generalized means, and design a provably fast and efficient approximation algorithm for the symmetrized centroid based on its exact geometric characterization that requires solely to walk on the geodesic linking the two sided centroids. We report on our generic implementation for computing entropic centers of image clusters and entropic centers of multivariate normals, and compare our results with former ad-hoc methods.

cs.CG↗

arXiv2007

Bregman Voronoi Diagrams: Properties, Algorithms and Applications

The Voronoi diagram of a finite set of objects is a fundamental geometric structure that subdivides the embedding space into regions, each region consisting of the points that are closer to a given object than to the others. We may define many variants of Voronoi diagrams depending on the class of objects, the distance functions and the embedding space. In this paper, we investigate a framework for defining and building Voronoi diagrams for a broad class of distance functions called Bregman divergences. Bregman divergences include not only the traditional (squared) Euclidean distance but also various divergence measures based on entropic functions. Accordingly, Bregman Voronoi diagrams allow to define information-theoretic Voronoi diagrams in statistical parametric spaces based on the relative entropy of distributions. We define several types of Bregman diagrams, establish correspondences between those diagrams (using the Legendre transformation), and show how to compute them efficiently. We also introduce extensions of these diagrams, e.g. k-order and k-bag Bregman Voronoi diagrams, and introduce Bregman triangulations of a set of points and their connexion with Bregman Voronoi diagrams. We show that these triangulations capture many of the properties of the celebrated Delaunay triangulation. Finally, we give some applications of Bregman Voronoi diagrams which are of interest in the context of computational geometry and machine learning.

cs.CG↗