Dissertationen zum Thema „Distances de Wasserstein“
Geben Sie eine Quelle nach APA, MLA, Chicago, Harvard und anderen Zitierweisen an
Machen Sie sich mit Top-49 Dissertationen für die Forschung zum Thema "Distances de Wasserstein" bekannt.
Neben jedem Werk im Literaturverzeichnis ist die Option "Zur Bibliographie hinzufügen" verfügbar. Nutzen Sie sie, wird Ihre bibliographische Angabe des gewählten Werkes nach der nötigen Zitierweise (APA, MLA, Harvard, Chicago, Vancouver usw.) automatisch gestaltet.
Sie können auch den vollen Text der wissenschaftlichen Publikation im PDF-Format herunterladen und eine Online-Annotation der Arbeit lesen, wenn die relevanten Parameter in den Metadaten verfügbar sind.
Sehen Sie die Dissertationen für verschiedene Spezialgebieten durch und erstellen Sie Ihre Bibliographie auf korrekte Weise.
Boissard, Emmanuel. „Problèmes d'interaction discret-continu et distances de Wasserstein“. Toulouse 3, 2011. http://thesesups.ups-tlse.fr/1389/.
Der volle Inhalt der QuelleWe study several problems of approximation using tools from Optimal Transportation theory. The family of Wasserstein metrics are used to provide error bounds for particular approximation of some Partial Differential Equations. They also come into play as natural measures of distorsion for quantization and clustering problems. A problem related to these questions is to estimate the speed of convergence in the empirical law of large numbers for these distorsions. The first part of this thesis provides non-asymptotic bounds, notably in infinite-dimensional Banach spaces, as well as in cases where independence is removed. The second part is dedicated to the study of two models from the modelling of animal displacement. A new individual-based model for ant trail formation is introduced, and studied through numerical simulations and kinetic formulation. We also study a variant of the Cucker-Smale model of bird flock motion : we establish well-posedness of the associated Vlasov-type transport equation as well as long-time behaviour results. In a third part, we study some statistical applications of the notion of barycenter in Wasserstein space recently introduced by M. Agueh and G. Carlier
Fernandes, Montesuma Eduardo. „Multi-Source Domain Adaptation through Wasserstein Barycenters“. Electronic Thesis or Diss., université Paris-Saclay, 2024. http://www.theses.fr/2024UPASG045.
Der volle Inhalt der QuelleMachine learning systems work under the assumption that training and test conditions are uniform, i.e., they do not change. However, this hypothesis is seldom met in practice. Hence, the system is trained with data that is no longer representative of the data it will be tested on. This case is represented by a shift in the probability measure generating the data. This scenario is known in the literature as distributional shift between two domains: a source, and a target. A straightforward generalization of this problem is when training data itself exhibit shifts on its own. In this case, one consider Multi Source Domain Adaptation (MSDA). In this context, optimal transport is an useful field of mathematics. Especially, optimal transport serves as a toolbox, for comparing and manipulating probability measures. This thesis studies the contributions of optimal transport to multi-source domain adaptation. We do so through Wasserstein barycenters, an object that defines a weighted average, in the space of probability measures, for the multiple domains in MSDA. Based on this concept, we propose: (i) a novel notion of barycenter, when the measures at hand are equipped with labels, (ii) a novel dictionary learning problem over empirical probability measures and (iii) new tools for domain adaptation through the optimal transport of Gaussian mixture models. Through our methods, we are able to improve domain adaptation performance in comparison with previous optimal transport-based methods on image, and cross-domain fault diagnosis benchmarks. Our work opens an interesting research direction, on learning the barycentric hull of probability measures
Schrieber, Jörn [Verfasser], Dominic [Akademischer Betreuer] Schuhmacher, Dominic [Gutachter] Schuhmacher und Anita [Gutachter] Schöbel. „Algorithms for Optimal Transport and Wasserstein Distances / Jörn Schrieber ; Gutachter: Dominic Schuhmacher, Anita Schöbel ; Betreuer: Dominic Schuhmacher“. Göttingen : Niedersächsische Staats- und Universitätsbibliothek Göttingen, 2019. http://d-nb.info/1179449304/34.
Der volle Inhalt der QuelleSEGUY, Vivien Pierre François. „Measure Transport Approaches for Data Visualization and Learning“. Kyoto University, 2018. http://hdl.handle.net/2433/233857.
Der volle Inhalt der QuelleGairing, Jan, Michael Högele, Tetiana Kosenkova und Alexei Kulik. „On the calibration of Lévy driven time series with coupling distances : an application in paleoclimate“. Universität Potsdam, 2014. http://opus.kobv.de/ubp/volltexte/2014/6978/.
Der volle Inhalt der QuelleFlenghi, Roberta. „Théorème de la limite centrale pour des fonctionnelles non linéaires de la mesure empirique et pour le rééchantillonnage stratifié“. Electronic Thesis or Diss., Marne-la-vallée, ENPC, 2023. http://www.theses.fr/2023ENPC0051.
Der volle Inhalt der QuelleThis thesis is dedicated to the central limit theorem which is one of the two fundamental limit theorems in probability theory with the strong law of large numbers.The central limit theorem which is well known for linear functionals of the empirical measure of independent and identically distributed random vectors, has recently been extended to non-linear functionals. The main tool permitting this extension is the linear functional derivative, one of the notions of derivation on the Wasserstein space of probability measures.We generalize this extension by first relaxing the equal distribution assumptionand then the independence property to be able to deal with the successive values of an ergodic Markov chain.In the second place, we focus on the stratified resampling mechanism.This is one of the resampling schemes commonly used in particle filters. We prove a central limit theorem for the first resampling according to this mechanism under the assumption that the initial positions are independent and identically distributed and the weights proportional to a positive function of the positions such that the image of their common distribution by this function has a non zero component absolutely continuous with respect to the Lebesgue measure. This result relies on the convergence in distribution of the fractional part of partial sums of the normalized weights to some random variable uniformly distributed on [0,1]. More generally, we prove the joint convergence in distribution of q variables modulo one obtained as partial sums of a sequence of i.i.d. square integrable random variables multiplied by a common factor given by some function of an empirical mean of the same sequence. The limit is uniformly distributed over [dollar][0,1]^q[dollar]. To deal with the coupling introduced by the common factor, we assume that the common distribution of the random variables has a non zero component absolutely continuous with respect to the Lebesgue measure, so that the convergence in the central limit theorem for this sequence holds in total variation distance.Under the conjecture that the convergence in distribution of fractional parts to some uniform random variable remains valid at the next steps of a particle filter which alternates selections according to the stratified resampling mechanism and mutations according to Markov kernels, we provide an inductive formula for the asymptotic variance of the resampled population after n steps. We perform numerical experiments which support the validity of this formula
Bobbia, Benjamin. „Régression quantile extrême : une approche par couplage et distance de Wasserstein“. Thesis, Bourgogne Franche-Comté, 2020. http://www.theses.fr/2020UBFCD043.
Der volle Inhalt der QuelleThis work is related with the estimation of conditional extreme quantiles. More precisely, we estimate high quantiles of a real distribution conditionally to the value of a covariate, potentially in high dimension. A such estimation is made introducing the proportional tail model. This model is studied with coupling methods. The first is an empirical processes based method whereas the second is focused on transport and optimal coupling. We provide estimators of both quantiles and model parameters, we show their asymptotic normality with our coupling methods. We also provide a validation procedure for proportional tail model. Moreover, we develop the second approach in the general framework of univariate extreme value theory
Nadjahi, Kimia. „Sliced-Wasserstein distance for large-scale machine learning : theory, methodology and extensions“. Electronic Thesis or Diss., Institut polytechnique de Paris, 2021. http://www.theses.fr/2021IPPAT050.
Der volle Inhalt der QuelleMany methods for statistical inference and generative modeling rely on a probability divergence to effectively compare two probability distributions. The Wasserstein distance, which emerges from optimal transport, has been an interesting choice, but suffers from computational and statistical limitations on large-scale settings. Several alternatives have then been proposed, including the Sliced-Wasserstein distance (SW), a metric that has been increasingly used in practice due to its computational benefits. However, there is little work regarding its theoretical properties. This thesis further explores the use of SW in modern statistical and machine learning problems, with a twofold objective: 1) provide new theoretical insights to understand in depth SW-based algorithms, and 2) design novel tools inspired by SW to improve its applicability and scalability. We first prove a set of asymptotic properties on the estimators obtained by minimizing SW, as well as a central limit theorem whose convergence rate is dimension-free. We also design a novel likelihood-free approximate inference method based on SW, which is theoretically grounded and scales well with the data size and dimension. Given that SW is commonly estimated with a simple Monte Carlo scheme, we then propose two approaches to alleviate the inefficiencies caused by the induced approximation error: on the one hand, we extend the definition of SW to introduce the Generalized Sliced-Wasserstein distances, and illustrate their advantages on generative modeling applications; on the other hand, we leverage concentration of measure results to formulate a new deterministic approximation for SW, which is computationally more efficient than the usual Monte Carlo technique and has nonasymptotical guarantees under a weak dependence condition. Finally, we define the general class of sliced probability divergences and investigate their topological and statistical properties; in particular, we establish that the sample complexity of any sliced divergence does not depend on the problem dimension
Liu, Lu. „A Risk-Oriented Clustering Approach for Asset Categorization and Risk Measurement“. Thesis, Université d'Ottawa / University of Ottawa, 2019. http://hdl.handle.net/10393/39444.
Der volle Inhalt der QuelleLescornel, Hélène. „Covariance estimation and study of models of deformations between distributions with the Wasserstein distance“. Toulouse 3, 2014. http://www.theses.fr/2014TOU30045.
Der volle Inhalt der QuelleThe first part of this thesis concerns the covariance estimation of non stationary processes. We are estimating the covariance in different vectorial spaces of matrices. In Chapter 3, we give a model selection procedure by minimizing a penalized criterion and using concentration inequalities, and Chapter 4 presents an Unbiased Risk Estimation method. In both cases we give oracle inequalities. The second part deals with the study of models of deformation between distributions. We assume that we observe a random quantity epsilon through a deformation function. The importance of the deformation is represented by a parameter theta that we aim to estimate. We present several methods of estimation based on the Wasserstein distance by aligning the distributions of the observations to recover the deformation parameter. In the case of real random variables, Chapter 7 presents properties of consistency for a M-estimator and its asymptotic distribution. We use Hadamard differentiability techniques to apply a functional Delta method. Chapter 8 concerns a Robbins-Monro estimator for the deformation parameter and presents properties of convergence for a kernel estimator of the density of the variable epsilon obtained with the observations. The model is generalized to random variables in complete metric spaces in Chapter 9. Then, in the aim to build a goodness of fit test, Chapter 10 gives results on the asymptotic distribution of a test statistic
Boistard, Hélène. „Eficacia asintotica tests relacionados con el estadística de Wasserstein“. Toulouse 3, 2007. http://www.theses.fr/2007TOU30155.
Der volle Inhalt der QuelleThe goodness of fit test based on the Wasserstein distance is a test which is well adapted to location-scale families. The asymptotic distribution under the null hypothesis has been known since the works by del Barrio et al. (1999, 2000). The subject of this thesis is the study of the asymptotic power of this test and of some related tests, owing to several efficiency criteria. In the first chapter, a short introduction presents the problem and the tools to be used. The second chapter is devoted to the the proof of some asymptotic results for multiple integrals with respect to the empirical process. These statistics are strongly related to U-statistics, but they permit an important simplification of the classical hypotheses to establish the asymptotic distribution under the null hypothesis, under contiguous alternative and for the bootstrap. In the third chapter, we prove that the Wasserstein test statistic is equivalent to a test based on the double integral with respect to the empirical process. This allows us to apply to this test the results of the previous chapter, and to obtain some information about its asymptotic efficiency in the framework of Gaussian shift experiments. .
Vidal, Jules. „Progressivité en analyse topologique de données“. Electronic Thesis or Diss., Sorbonne université, 2021. http://www.theses.fr/2021SORUS398.
Der volle Inhalt der QuelleTopological Data Analysis (TDA) forms a collection of tools that enable the generic and efficient extraction of features in data. However, although most TDA algorithms have practicable asymptotic complexities, these methods are rarely interactive on real-life datasets, which limits their usability for interactive data analysis and visualization. In this thesis, we aimed at developing progressive methods for the TDA of scientific scalar data, that can be interrupted to swiftly provide a meaningful approximate output and that are able to refine it otherwise. First, we introduce two progressive algorithms for the computation of the critical points and the extremum-saddle persistence diagram of a scalar field. Next, we revisit this progressive framework to introduce an approximation algorithm for the persistence diagram of a scalar field, with strong guarantees on the related approximation error. Finally, in a effort to perform visual analysis of ensemble data, we present a novel progressive algorithm for the computation of the discrete Wasserstein barycenter of a set of persistence diagrams, a notoriously computationally intensive task. Our progressive approach enables the approximation of the barycenter within interactive times. We extend this method to a progressive, time-constraint, topological ensemble clustering algorithm
Lebrat, Léo. „Projection au sens de Wasserstein 2 sur des espaces structurés de mesures“. Thesis, Toulouse, INSA, 2019. http://www.theses.fr/2019ISAT0035.
Der volle Inhalt der QuelleThis thesis focuses on the approximation for the 2-Wasserstein metric of probability measures by structured measures. The set of structured measures under consideration is made of consistent discretizations of measures carried by a smooth curve with a bounded speed and acceleration. We compare two different types of approximations of the curve: piecewise constant and piecewise linear. For each of these methods, we develop fast and scalable algorithms to compute the 2-Wasserstein distance between a given measure and the structured measure. The optimization procedure reveals new theoretical and numerical challenges, it consists of two steps: first the computation of the 2-Wasserstein distance, second the optimization of the parameters of structure. This work is initially motivated by the design of trajectories in MRI acquisition, however we provide new applications of these methods
Sommerfeld, Max [Verfasser], Axel [Akademischer Betreuer] Munk, Axel [Gutachter] Munk und Stephan [Gutachter] Huckemann. „Wasserstein Distance on Finite Spaces: Statistical Inference and Algorithms / Max Sommerfeld ; Gutachter: Axel Munk, Stephan Huckemann ; Betreuer: Axel Munk“. Göttingen : Niedersächsische Staats- und Universitätsbibliothek Göttingen, 2017. http://d-nb.info/1149959223/34.
Der volle Inhalt der QuelleBuzun, Nazar. „Bootstrap in high dimensional spaces“. Doctoral thesis, Humboldt-Universität zu Berlin, 2021. http://dx.doi.org/10.18452/22285.
Der volle Inhalt der QuelleThe objective of this thesis is to explore theoretical properties of various bootstrap methods. We introduce the convergence rates of the bootstrap procedure which corresponds to the difference between real distribution of some statistic and its resampling approximation. In this work we analyze the distribution of Euclidean norm of independent vectors sum, maximum of sum in high dimension, Wasserstein distance between empirical measures, Wassestein barycenters. In order to prove bootstrap convergence we involve Gaussian approximation technique which means that one has to find a sum of independent vectors in the considered statistic such that bootstrap yields a resampling of this sum. Further this sum may be approximated by Gaussian distribution and compared with the resampling distribution as a difference between variance matrices. In general it appears to be very difficult to reveal such a sum of independent vectors because some statistics (for example, MLE) don't have an explicit equation and may be infinite-dimensional. In order to handle this difficulty we involve some novel results from statistical learning theory, which provide a finite sample quadratic approximation of the Likelihood and suitable MLE representation. In the last chapter we consider the MLE of Wasserstein barycenters model. The regularised barycenters model has bounded derivatives and satisfies the necessary conditions of quadratic approximation. Furthermore, we apply bootstrap in change point detection methods. In the parametric case we analyse the Likelihood Ratio Test (LRT) statistic. Its high values indicate changes of parametric distribution in the data sequence. The maximum of LRT has a complex distribution but its quantiles may be calibrated by means of bootstrap. We show the convergence rates of the bootstrap quantiles to the real quantiles of LRT distribution. In non-parametric case instead of LRT we use Wasserstein distance between empirical measures. We test the accuracy of change point detection methods on synthetic time series and electrocardiography (ECG) data. Experiments with ECG illustrate advantages of the non-parametric approach versus complex parametric models and LRT.
Durmus, Alain. „High dimensional Markov chain Monte Carlo methods : theory, methods and applications“. Thesis, Université Paris-Saclay (ComUE), 2016. http://www.theses.fr/2016SACLT001/document.
Der volle Inhalt der QuelleThe subject of this thesis is the analysis of Markov Chain Monte Carlo (MCMC) methods and the development of new methodologies to sample from a high dimensional distribution. Our work is divided into three main topics. The first problem addressed in this manuscript is the convergence of Markov chains in Wasserstein distance. Geometric and sub-geometric convergence with explicit constants, are derived under appropriate conditions. These results are then applied to thestudy of MCMC algorithms. The first analyzed algorithm is an alternative scheme to the Metropolis Adjusted Langevin algorithm for which explicit geometric convergence bounds are established. The second method is the pre-Conditioned Crank-Nicolson algorithm. It is shown that under mild assumption, the Markov chain associated with thisalgorithm is sub-geometrically ergodic in an appropriated Wasserstein distance. The second topic of this thesis is the study of the Unadjusted Langevin algorithm (ULA). We are first interested in explicit convergence bounds in total variation under different kinds of assumption on the potential associated with the target distribution. In particular, we pay attention to the dependence of the algorithm on the dimension of the state space. The case of fixed step sizes as well as the case of nonincreasing sequences of step sizes are dealt with. When the target density is strongly log-concave, explicit bounds in Wasserstein distance are established. These results are then used to derived new bounds in the total variation distance which improve the one previously derived under weaker conditions on the target density.The last part tackles new optimal scaling results for Metropolis-Hastings type algorithms. First, we extend the pioneer result on the optimal scaling of the random walk Metropolis algorithm to target densities which are differentiable in Lp mean for p ≥ 2. Then, we derive new Metropolis-Hastings type algorithms which have a better optimal scaling compared the MALA algorithm. Finally, the stability and the convergence in total variation of these new algorithms are studied
Feyeux, Nelson. „Transport optimal pour l'assimilation de données images“. Thesis, Université Grenoble Alpes (ComUE), 2016. http://www.theses.fr/2016GREAM076/document.
Der volle Inhalt der QuelleForecasting of a physical system is computed by the help of a mathematical model. This model needs to be initialized by the state of the system at initial time. But this state is not directly measurable and data assimilation techniques are generally used to estimate it. They combine all sources of information such as observations (that may be sparse in time and space and potentially include errors), previous forecasts, the model equations and error statistics. The main idea of data assimilation techniques is to find an initial state accounting for the different sources of informations. Such techniques are widely used in meteorology, where data and particularly images are more and more numerous due to the increasing number of satellites and other sources of measurements. This, coupled with developments of meteorological models, have led to an ever-increasing quality of the forecast.Spatial consistency is one specificity of images. For example, human eyes are able to notice structures in an image. However, classical methods of data assimilation do not handle such structures because they take only into account the values of each pixel separately. In some cases it leads to a bad initial condition. To tackle this problem, we proposed to change the representation of an image: images are considered here as elements of the Wasserstein space endowed with the Wasserstein distance coming from the optimal transport theory. In this space, what matters is the positions of the different structures.This thesis presents a data assimilation technique based on this Wasserstein distance. This technique and its numerical procedure are first described, then experiments are carried out and results shown. In particularly, it appears that this technique was able to give an analysis of corrected position
Cloez, Bertrand. „Comportement asymptotique de processus avec sauts et applications pour des modèles avec branchement“. Phd thesis, Université Paris-Est, 2013. http://tel.archives-ouvertes.fr/tel-00862913.
Der volle Inhalt der QuelleLaborde, Maxime. „Systèmes de particules en interaction, approche par flot de gradient dans l'espace de Wasserstein“. Thesis, Paris Sciences et Lettres (ComUE), 2016. http://www.theses.fr/2016PSLED014/document.
Der volle Inhalt der QuelleSince 1998 and the seminal work of Jordan, Kinderlehrer and Otto, it is well known that a large class of parabolic equations can be seen as gradient flows in the Wasserstein space. This thesis is devoted to extensions of this theory to equations and systems which do not have exactly a gradient flow structure. We study different kind of couplings. First, we treat the case of nonlocal interactions in the drift. Then, we study cross diffusion systems which model congestion for several species. We are also interested in reaction-diffusion systems as diffusive prey-predator systems or tumor growth models. Finally, we introduce a new class of systems where the interaction is given by a multi-marginal transport problem. In many cases, we give numerical simulations to illustrate our theorical results
Christen, Alejandra. „Comportamiento asintótico de los procesos de Markov deterministas por pedazos“. Tesis, Universidad de Chile, 2012. http://www.repositorio.uchile.cl/handle/2250/111925.
Der volle Inhalt der QuelleEn esta tesis doctoral se abordan dos problemas relacionados con el comportamiento en tiempo largo de los procesos de Markov deterministas por pedazos (PDMP). En primer lugar se estudia el comportamiento asintótico de un PDMP general en relación con el comportamiento y propiedades de una cadena de Markov a tiempo discreto embuída. Este problema se desarrolla en el Capítulo 1. En segundo lugar, se considera un PDMP específico llamado Proceso del tamaño de ventana del TCP (sigla en inglés del protocolo de control de transmisión usado en internet). El objetivo en este caso es encontrar tasas de convergencia explícitas al equilibrio. Este problema se estudia en el Capítulo 2. Con respecto al primer problema, en el Cap´ıtulo 1 se relacionan las propiedades de recurrencia positiva y las medidas de probabilidad invariantes de un proceso PDMP general con las de una cadena espacio-tiempo discreta, formada por las posiciones post-salto del proceso y las longitudes de tiempo entre saltos. Esta cadena discreta se obtiene de manera simple a partir de las características locales que definen el proceso a tiempo continuo y contiene más información que la cadena discreta post-salto que ha sido habitualmente considerada. Utilizando esta cadena espacio-tiempo se puede definir un nuevo proceso a tiempo continuo asociado, formado por tres coordenadas: el proceso continuo propiamente dicho, la longitud de tiempo trancurrido desde el último tiempo de salto y la longitud de tiempo que falta para el siguiente tiempo de salto, en analogía con los procesos edad y vida residual de teoría de renovación. En este capítulo se describe completamente el equilibrio de este proceso asociado y se establece un resultado análogo de la waiting time paradox de teoría de renovación en el contexto de los PDMP. Para el segundo problema, en el Capítulo 2 se obtienen tasas de convergencia exponencial al equilibrio en distancia Wasserstein y en la norma en variación total. Estos resultados se basan en algunos argumentos de acoplamiento nuevos y dan una respuesta a una pregunta importante sobre el protocolo de transmisión de internet TCP, que es el entender cómo la congestión del tamaño de ventana del TCP alcanza equilibrio en tiempo largo.
Kampschulte, Malte [Verfasser], Christof Erich [Akademischer Betreuer] Melcher, Michael [Akademischer Betreuer] Westdickenberg und Robert L. [Akademischer Betreuer] Jerrard. „Gradient flows and a generalized Wasserstein distance in the space of Cartesian currents / Malte Kampschulte ; Christof Erich Melcher, Michael Westdickenberg, Robert L. Jerrard“. Aachen : Universitätsbibliothek der RWTH Aachen, 2017. http://d-nb.info/1187346314/34.
Der volle Inhalt der QuelleTinarrage, Raphaël. „Inférence topologique à partir de mesures et de fibrés vectoriels“. Thesis, université Paris-Saclay, 2020. http://www.theses.fr/2020UPASM001.
Der volle Inhalt der QuelleWe contribute to the theory of topological inference, based on the theory of persistent homology, by proposing three families of filtrations.For each of them, we prove consistency results---that is, the quality of approximation of an underlying geometric object---, and stability results---that is, robustness against initial measurement errors.We propose concrete algorithms in order to use these methods in practice.The first family, the DTM-filtration, is a robust alternative to the classical Cech filtration when the point cloud is noisy or contains outliers.It is based on the notion of distance to measure, which allows to obtain stability in the sense of the Wasserstein distance.Secondly, we propose the lifted filtrations, which make it possible to estimate the homology of immersed manifolds, even when their reach is zero.We introduce the notion of normal reach, and show that it leads to a quantitative control of the manifold.We study the estimation of tangent spaces by local covariance matrices.Thirdly, we develop a framework for vector bundle filtrations, and define the persistent Stiefel-Whitney classes.We show that the persistent classes associated to the Cech bundle filtrations are Hausdorff-stable and consistent.To allow their algorithmic implementation, we introduce the notion of weak star condition
Durmus, Alain. „High dimensional Markov chain Monte Carlo methods : theory, methods and applications“. Electronic Thesis or Diss., Université Paris-Saclay (ComUE), 2016. http://www.theses.fr/2016SACLT001.
Der volle Inhalt der QuelleThe subject of this thesis is the analysis of Markov Chain Monte Carlo (MCMC) methods and the development of new methodologies to sample from a high dimensional distribution. Our work is divided into three main topics. The first problem addressed in this manuscript is the convergence of Markov chains in Wasserstein distance. Geometric and sub-geometric convergence with explicit constants, are derived under appropriate conditions. These results are then applied to thestudy of MCMC algorithms. The first analyzed algorithm is an alternative scheme to the Metropolis Adjusted Langevin algorithm for which explicit geometric convergence bounds are established. The second method is the pre-Conditioned Crank-Nicolson algorithm. It is shown that under mild assumption, the Markov chain associated with thisalgorithm is sub-geometrically ergodic in an appropriated Wasserstein distance. The second topic of this thesis is the study of the Unadjusted Langevin algorithm (ULA). We are first interested in explicit convergence bounds in total variation under different kinds of assumption on the potential associated with the target distribution. In particular, we pay attention to the dependence of the algorithm on the dimension of the state space. The case of fixed step sizes as well as the case of nonincreasing sequences of step sizes are dealt with. When the target density is strongly log-concave, explicit bounds in Wasserstein distance are established. These results are then used to derived new bounds in the total variation distance which improve the one previously derived under weaker conditions on the target density.The last part tackles new optimal scaling results for Metropolis-Hastings type algorithms. First, we extend the pioneer result on the optimal scaling of the random walk Metropolis algorithm to target densities which are differentiable in Lp mean for p ≥ 2. Then, we derive new Metropolis-Hastings type algorithms which have a better optimal scaling compared the MALA algorithm. Finally, the stability and the convergence in total variation of these new algorithms are studied
Guittet, Kévin. „Contributions à la résolution numérique de problèmes de transport optimal de masse“. Paris 6, 2003. http://www.theses.fr/2003PA066380.
Der volle Inhalt der QuelleSalem, Samir. „Limite de champ moyen et propagation du chaos pour des systèmes de particules avec interaction discontinue“. Thesis, Aix-Marseille, 2017. http://www.theses.fr/2017AIXM0387/document.
Der volle Inhalt der QuelleIn this thesis, we study some propagation of chaos and mean field limit problems arising in modelisation of collective behavior of individuals or particles. Particularly, we set ourselves in the case where the interaction between the individuals/particles is discontinuous. The first work establihes the propagation of chaos for the 1d Vlasov-Poisson-Fokker-Planck equation. More precisely, we show that the distribution of particles evolving on the real line and interacting through the sign function converges to the solution of the 1d VPFP equation, in probability by large deviations-like techniques, and in expectation by law of large numbers-like techniques. In the second work, we study a variant of the Cucker-Smale, where the communication weight is the indicatrix function of a cone which orientation depends on the velocity of the individual. Some weak-strong stability estimate in M.K.W. distance is obtained for the limit equation, which implies the mean field limit. The third work consists in adding some diffusion in velocity to the model previously quoted. However one must add some truncated diffusion in order to preserve a system in which velocities remain unifomrly bounded. Finally we study a variant of the aggregation equation where the interaction between individuals is also given by a cone which orientation depends on the position of the individual. In this case we are only able to provide some weak-strong stability estimate in $W_\infty$ distance, and the problem must be set in a bounded domain for the case with diffusion
Margheriti, William. „Sur la stabilité du problème de transport optimal martingale“. Thesis, Paris Est, 2020. http://www.theses.fr/2020PESC1036.
Der volle Inhalt der QuelleThis thesis is motivated by the study of the stability of the Martingale Optimal Transport problem, and is naturally structured around two parts. In the first part, we exhibit a new family of martingale couplings between two one-dimensional probability measures μ and ν in the convex order. This family contains in particular the inverse transform martingale coupling which is explicit in terms of the quantile functions of these marginal densities. The integral M_1(μ,ν) of |x-y| with respect to each of these couplings is smaller than or equal to twice the Wasserstein distance W_1(μ,ν) between μ and ν. We show that a similar inequality holds when replacing |x-y| and W_1 respectively with |x-y|^ρ and the product of W_ρ times the centered ρ-th moment of the second marginal to the power ρ-1, for any ρ∈[1,+∞). We then study the generalisation of this new stability inequality to higher dimensions. Last, we establish a strong connection between our new family of martingale couplings and the projection of a coupling between two given marginals in the convex order onto the set of martingale couplings between the same marginals. The latter projection is taken with respect to the adapted Wasserstein distance, which is greater than or equal to the usual Wasserstein distance and therefore induces a finer topology, which is more suitable to financial modelisation since it takes into account the temporal structure of martingales. In the second part, we prove that any martingale coupling whose marginals are approximated by probability measures in the convex order can be approximated by martingale couplings with respected to the adapted Wasserstein distance. We then discuss various applications of this result. In particular, we strengthen a stability result on the Optimal Weak Transport problem and establish a stability result on the Martingale Optimal Weak Transport problem. We deduce the stability with respect to the marginals of the superreplication price of VIX futures
Arsenteva, Polina. „Statistical modeling and analysis of radio-induced adverse effects based on in vitro and in vivo data“. Electronic Thesis or Diss., Bourgogne Franche-Comté, 2023. http://www.theses.fr/2023UBFCK074.
Der volle Inhalt der QuelleIn this work we address the problem of adverse effects induced by radiotherapy on healthy tissues. The goal is to propose a mathematical framework to compare the effects of different irradiation modalities, to be able to ultimately choose those treatments that produce the minimal amounts of adverse effects for potential use in the clinical setting. The adverse effects are studied in the context of two types of data: in terms of the in vitro omic response of human endothelial cells, and in terms of the adverse effects observed on mice in the framework of in vivo experiments. In the in vitro setting, we encounter the problem of extracting key information from complex temporal data that cannot be treated with the methods available in literature. We model the radio-induced fold change, the object that encodes the difference in the effect of two experimental conditions, in the way that allows to take into account the uncertainties of measurements as well as the correlations between the observed entities. We construct a distance, with a further generalization to a dissimilarity measure, allowing to compare the fold changes in terms of all the important statistical properties. Finally, we propose a computationally efficient algorithm performing clustering jointly with temporal alignment of the fold changes. The key features extracted through the latter are visualized using two types of network representations, for the purpose of facilitating biological interpretation. In the in vivo setting, the statistical challenge is to establish a predictive link between variables that, due to the specificities of the experimental design, can never be observed on the same animals. In the context of not having access to joint distributions, we leverage the additional information on the observed groups to infer the linear regression model. We propose two estimators of the regression parameters, one based on the method of moments and the other based on optimal transport, as well as the estimators for the confidence intervals based on the stratified bootstrap procedure
Gordaliza, Pastor Paula. „Fair learning : une approche basée sur le transport optimale“. Thesis, Toulouse 3, 2020. http://www.theses.fr/2020TOU30084.
Der volle Inhalt der QuelleThe aim of this thesis is two-fold. On the one hand, optimal transportation methods are studied for statistical inference purposes. On the other hand, the recent problem of fair learning is addressed through the prism of optimal transport theory. The generalization of applications based on machine learning models in the everyday life and the professional world has been accompanied by concerns about the ethical issues that may arise from the adoption of these technologies. In the first part of the thesis, we motivate the fairness problem by presenting some comprehensive results from the study of the statistical parity criterion through the analysis of the disparate impact index on the real and well-known Adult Income dataset. Importantly, we show that trying to make fair machine learning models may be a particularly challenging task, especially when the training observations contain bias. Then a review of Mathematics for fairness in machine learning is given in a general setting, with some novel contributions in the analysis of the price for fairness in regression and classification. In the latter, we finish this first part by recasting the links between fairness and predictability in terms of probability metrics. We analyze repair methods based on mapping conditional distributions to the Wasserstein barycenter. Finally, we propose a random repair which yields a tradeoff between minimal information loss and a certain amount of fairness. The second part is devoted to the asymptotic theory of the empirical transportation cost. We provide a Central Limit Theorem for the Monge-Kantorovich distance between two empirical distributions with different sizes n and m, Wp(Pn,Qm), p > = 1, for observations on R. In the case p > 1 our assumptions are sharp in terms of moments and smoothness. We prove results dealing with the choice of centering constants. We provide a consistent estimate of the asymptotic variance which enables to build two sample tests and confidence intervals to certify the similarity between two distributions. These are then used to assess a new criterion of data set fairness in classification. Additionally, we provide a moderate deviation principle for the empirical transportation cost in general dimension. Finally, Wasserstein barycenters and variance-like criterion using Wasserstein distance are used in many problems to analyze the homogeneity of collections of distributions and structural relationships between the observations. We propose the estimation of the quantiles of the empirical process of the Wasserstein's variation using a bootstrap procedure. Then we use these results for statistical inference on a distribution registration model for general deformation functions. The tests are based on the variance of the distributions with respect to their Wasserstein's barycenters for which we prove central limit theorems, including bootstrap versions
Gigli, Nicola. „Proprietes geometriques et analytiques de certaines structures non lisses“. Habilitation à diriger des recherches, Université de Nice Sophia-Antipolis, 2011. http://tel.archives-ouvertes.fr/tel-00769381.
Der volle Inhalt der QuelleGiorgi, Pierre-Antoine. „Analyse mathématique de modèles cinétiques en physique des plasmas“. Electronic Thesis or Diss., Aix-Marseille, 2019. http://www.theses.fr/2019AIXM0609.
Der volle Inhalt der QuelleThis thesis deals with the study of some kinetic models encountered in plasma physics.The first model considered is a 1D Vlasov-Poisson system representing the dynamics of two species of particles (ions and electrons) in a bounded set, x ∈ (0,1), with direct reflection boundary conditions. In the linear case, generalized characteristics are defined, ensuring the time s=0 to be reached after a finite number of bounces, the problematic case being when the electric field points outward of the boundary. Then, for initial conditions even in the velocity variable, a global continuous solution is built by means of generalized characteristics and a fixed point argument. Local uniqueness of a continuous solution is shown, in a frame where two successive bounces at the same boundary cannot occur. The second model was obtained as the limit of a Vlasov-Poisson system in the finite Larmor radius regime.For solutions satisfying a decay assumption, a Wasserstein stability estimate is proven, and a new proof of the existence of such solutions is given. The advection field is then Lipschitz continuous. Finally, numerical simulations are performed to investigate the kinetic response of electrons to an external drive. A beating between two waves, one at the external frequency, the other at the Landau frequency, is revealed
Zhang, Chaoen. „Long time behaviour of kinetic equations“. Thesis, Université Clermont Auvergne (2017-2020), 2019. http://www.theses.fr/2019CLFAC056.
Der volle Inhalt der QuelleThis dissertation is devoted to the long time behaviour of the kinetic Fokker-Planck equation and of the McKean-Vlasov equation. The manuscript is composed of an introduction and six chapters.The kinetic Fokker-Planck equation is a basic example for Villani's hypocoercivity theory which asserts the exponential decay in large time in the absence of coercivity. In his memoir, Villani proved the hypocoercivity for the kinetic Fokker-Planck equation in either weighted H^1, weighted L^2 or entropy.However, a boundedness condition of the Hessian of the Hamiltonian was imposed in the entropic case. We show in Chapter 2 how we can get rid of this assumption by well-chosen multipliers with the help of a weighted logarithmic Sobolev inequality. Such a functional inequality can be obtained by some tractable Lyapunov condition.In Chapter 4, we apply Villani's ideas and some Lyapunov conditions to prove hypocoercivity in weighted H^1 in the case of mean-field interaction with a rate of exponential convergence independent of the number N of particles. For proving this we should prove the Poincaré inequality with a constant independent of N, and rends a dimension dependent boundeness estimate of Villani dimension-free by means of the stronger uniform log-Sobolev inequality and Lyapunov function method. In Chapter 6, we study the hypocoercive contraction in L^2-Wasserstein distance and we recover the optimal rate in the quadratic potential case. The method is based on the temporal derivative of the Wasserstein distance.In Chapter 7, Villani's hypoercivity theorem in weighted H^1 space is extended to weighted H^k spaces by choosing carefully some appropriate mixed terms in the definition of norm of H^k.The McKean-Vlasov equation is a nonlinear nonlocal diffusive equation. It is well-Known that it has a gradient flow structure. However, the known results strongly depend on convexity assumptions. Such assumptions are notably relaxed in Chapter 3 and Chapter 5 where we prove the exponential convergence to equilibrium respectively in free energy and the L^1-Wasserstain distance. Our approach is based on the mean field limit theory. That is, we study the associated system of a large numer of paricles with mean-field interaction and then pass to the limit by propagation of chaos
Bui, Thi Thien Trang. „Modèle de régression pour des données non-Euclidiennes en grande dimension. Application à la classification de taxons en anatomie computationnelle“. Thesis, Toulouse, INSA, 2019. http://www.theses.fr/2019ISAT0021.
Der volle Inhalt der QuelleIn this thesis, we study a regression model with distribution entries and the testing hypothesis problem for signal detection in a regression model. We aim to apply these models in hearing sensitivity measured by the transient evoked otoacoustic emissions (TEOAEs) data to improve our knowledge in the auditory investigation. In the first part, a new distribution regression model for probability distributions is introduced. This model is based on a Reproducing Kernel Hilbert Space (RKHS) regression framework, where universal kernels are built using Wasserstein distances for distributions belonging to \Omega) and \Omega is a compact subspace of the real space. We prove the universal kernel property of such kernels and use this setting to perform regressions on functions. Different regression models are first compared with the proposed one on simulated functional data. We then apply our regression model to transient evoked otoascoutic emission (TEOAE) distribution responses and real predictors of the age. This part is a joint work with Loubes, J-M., Risser, L. and Balaresque, P..In the second part, considering a regression model, we address the question of testing the nullity of the regression function. The testing procedure is available when the variance of the observations is unknown and does not depend on any prior information on the alternative. We first propose a single testing procedure based on a general symmetric kernel and an estimation of the variance of the observations. The corresponding critical values are constructed to obtain non asymptotic level \alpha tests. We then introduce an aggregation procedure to avoid the difficult choice of the kernel and of the parameters of the kernel. The multiple tests satisfy non-asymptotic properties and are adaptive in the minimax sense over several classes of regular alternatives
Bonnotte, Nicolas. „Unidimensional and Evolution Methods for Optimal Transportation“. Phd thesis, Université Paris Sud - Paris XI, 2013. http://tel.archives-ouvertes.fr/tel-00946781.
Der volle Inhalt der QuelleMalrieu, Florent. „Inégalités fonctionnelles et comportement en temps long de quelques processus de Markov“. Habilitation à diriger des recherches, Université Rennes 1, 2010. http://tel.archives-ouvertes.fr/tel-00542278.
Der volle Inhalt der QuelleLiero, Matthias. „Variational methods for evolution“. Doctoral thesis, Humboldt-Universität zu Berlin, Mathematisch-Naturwissenschaftliche Fakultät II, 2013. http://dx.doi.org/10.18452/16685.
Der volle Inhalt der QuelleThis thesis deals with the application of variational methods to evolution problems governed by partial differential equations. The first part of this work is devoted to systems of reaction-diffusion equations that can be formulated as gradient systems with respect to an entropy functional and a dissipation metric. We provide methods for establishing geodesic convexity of the entropy functional by purely differential methods. Geodesic convexity is beneficial, however, it is a strong structural property of a gradient system that is rather difficult to achieve. Several examples, including a drift-diffusion system, provide a survey on the applicability of the theory. Next, we demonstrate the application of Gamma-convergence, to derive effective limit models for multiscale problems. The crucial point in this investigation is that we rely only on the gradient structure of the systems. We consider two model problems: The rigorous derivation of an Allen-Cahn system with bulk/surface coupling and of an interface condition for a one-dimensional diffusion equation. The second part of this thesis is devoted to the so-called Weighted-Inertia-Dissipation-Energy principle. The WIDE principle is a global-in-time variational principle for evolution equations either of conservative or dissipative type. It relies on the minimization of a specific parameter-dependent family of functionals (WIDE functionals) with minimizers characterizing entire trajectories of the system. We prove that minimizers of the WIDE functional converge, up to subsequences, to weak solutions of the limiting PDE when the parameter tends to zero. The interest for this perspective is that of moving the successful machinery of the Calculus of Variations.
Riou-Durand, Lionel. „Theoretical contributions to Monte Carlo methods, and applications to Statistics“. Thesis, Université Paris-Saclay (ComUE), 2019. http://www.theses.fr/2019SACLG006/document.
Der volle Inhalt der QuelleThe first part of this thesis concerns the inference of un-normalized statistical models. We study two methods of inference based on sampling, known as Monte-Carlo MLE (Geyer, 1994), and Noise Contrastive Estimation (Gutmann and Hyvarinen, 2010). The latter method was supported by numerical evidence of improved stability, but no theoretical results had yet been proven. We prove that Noise Contrastive Estimation is more robust to the choice of the sampling distribution. We assess the gain of accuracy depending on the computational budget. The second part of this thesis concerns approximate sampling for high dimensional distributions. The performance of most samplers deteriorates fast when the dimension increases, but several methods have proven their effectiveness (e.g. Hamiltonian Monte Carlo, Langevin Monte Carlo). In the continuity of some recent works (Eberle et al., 2017; Cheng et al., 2018), we study some discretizations of the kinetic Langevin diffusion process and establish explicit rates of convergence towards the sampling distribution, that scales polynomially fast when the dimension increases. Our work improves and extends the results established by Cheng et al. for log-concave densities
Reygner, Julien. „Comportements en temps long et à grande échelle de quelques dynamiques de collision“. Thesis, Paris 6, 2014. http://www.theses.fr/2014PA066471/document.
Der volle Inhalt der QuelleThis thesis contains three independent parts, each one of which is dedicated to the study of a particle system, following either a deterministic or a stochastic dynamics, and in which interactions only occur at collisions. Part I contains a numerical and theoretical study of nonequilibrium steady states of the Complete Exchange Model, which was introduced by physicists in order to understand heat transfer in some porous materials. Part II is dedicated to a system of Brownian particles evolving on the real line and interacting through their ranks. The long time and mean-field behaviour of this system is described, then the results are applied to the study of a model of equity market called the mean-field Atlas model. Part III introduces a multitype version of the particle system studied in the previous part, which allows to approximate parabolic systems of nonlinear partial differential equations. The small noise limit of of this system is called multitype sticky particle dynamics and now approximates hyperbolic systems. A detailed study of this dynamics provides stability estimates in Wasserstein distance for the solutions of these systems
Kashlak, Adam B. „A concentration inequality based statistical methodology for inference on covariance matrices and operators“. Thesis, University of Cambridge, 2017. https://www.repository.cam.ac.uk/handle/1810/267833.
Der volle Inhalt der QuelleRiou-Durand, Lionel. „Theoretical contributions to Monte Carlo methods, and applications to Statistics“. Electronic Thesis or Diss., Université Paris-Saclay (ComUE), 2019. http://www.theses.fr/2019SACLG006.
Der volle Inhalt der QuelleThe first part of this thesis concerns the inference of un-normalized statistical models. We study two methods of inference based on sampling, known as Monte-Carlo MLE (Geyer, 1994), and Noise Contrastive Estimation (Gutmann and Hyvarinen, 2010). The latter method was supported by numerical evidence of improved stability, but no theoretical results had yet been proven. We prove that Noise Contrastive Estimation is more robust to the choice of the sampling distribution. We assess the gain of accuracy depending on the computational budget. The second part of this thesis concerns approximate sampling for high dimensional distributions. The performance of most samplers deteriorates fast when the dimension increases, but several methods have proven their effectiveness (e.g. Hamiltonian Monte Carlo, Langevin Monte Carlo). In the continuity of some recent works (Eberle et al., 2017; Cheng et al., 2018), we study some discretizations of the kinetic Langevin diffusion process and establish explicit rates of convergence towards the sampling distribution, that scales polynomially fast when the dimension increases. Our work improves and extends the results established by Cheng et al. for log-concave densities
Godinho, David. „Contribution à l'étude des équations de Boltzmann, Kac et Keller-Segel à l'aide d'équations différentielles stochastiques non linéaires“. Phd thesis, Université Paris-Est, 2013. http://tel.archives-ouvertes.fr/tel-00931392.
Der volle Inhalt der QuelleSchrieber, Jörn. „Algorithms for Optimal Transport and Wasserstein Distances“. Doctoral thesis, 2019. http://hdl.handle.net/11858/00-1735-0000-002E-E5B2-B.
Der volle Inhalt der QuelleSommerfeld, Max. „Wasserstein Distance on Finite Spaces: Statistical Inference and Algorithms“. Doctoral thesis, 2017. http://hdl.handle.net/11858/00-1735-0000-0023-3FA1-C.
Der volle Inhalt der QuelleLUINI, EDOARDO GLAUCO. „Nonparametric density estimation with Wasserstein distance for actuarial applications“. Doctoral thesis, 2020. http://hdl.handle.net/11573/1365683.
Der volle Inhalt der QuelleHSU, WEI-CHE, und 許維哲. „A Generative Adversarial Network in Domain Adaptation by Utilizing the Wasserstein distance“. Thesis, 2019. http://ndltd.ncl.edu.tw/handle/nuv37w.
Der volle Inhalt der Quelle國立中正大學
電機工程研究所
107
Transfer learning is an important branch of machine learning, mainly used with unsupervised learning or semi-unsupervised learning, these two traditional machine learning methods. Learning important features of prior knowledge domain and generalizing them into different domains is the main subject. Domain adaptation is an important field of transfer learning and its the main goal is to measure the distribution of source domain as input and target domain output to increase the classification accuracy in unlabeled data. The powerful adversarial ability of generating GAN (Generative Adversarial Nets, GAN) makes the generated pictures more similar to real pictures, and derives a variety of GAN variants. In order to reduce training cost and rapidly transfer data in the domain adaptation, the main contribution in our research is applying MLP (Multi-layer perceptron) as the main component architecture, and using Wasserstein distance to evaluate the distribution of the probability between source and the target data in domain adaptation. We improve some disadvantages of GAN by using Consistency Term to modify the penalty function. Another contribution in our research is to apply our architecture in handwritten recognition dataset (MNIST, USPS) and object recognition dataset (Office-Caltech10) with two different features (SURF, DeCAF), and compare different deep transfer learning methods with 12 tasks between the source and target.
„Transportation Techniques for Geometric Clustering“. Doctoral diss., 2020. http://hdl.handle.net/2286/R.I.57239.
Der volle Inhalt der QuelleDissertation/Thesis
Doctoral Dissertation Computer Engineering 2020
Tameling, Carla. „Empirical Optimal Transport on Discrete Spaces: Limit Theorems, Distributional Bounds and Applications“. Doctoral thesis, 2018. http://hdl.handle.net/11858/00-1735-0000-002E-E552-1.
Der volle Inhalt der QuelleJabłoński, Jędrzej. „Structured population models for predator-prey interactions. The case of Daphnia and size selective planktivorous fish“. Doctoral thesis, 2014.
Den vollen Inhalt der Quelle findenIACOMINI, ELISA. „Mathematical models and methods for traffic flow and stop & go waves“. Doctoral thesis, 2020. http://hdl.handle.net/11573/1387448.
Der volle Inhalt der QuelleCAPONERA, ALESSIA. „Statistical inference for spherical functional autoregressions“. Doctoral thesis, 2020. http://hdl.handle.net/11573/1363165.
Der volle Inhalt der Quelle