Indice
Letteratura scientifica selezionata sul tema "Principe de minimisation du risque empirique"
Cita una fonte nei formati APA, MLA, Chicago, Harvard e in molti altri stili
Consulta la lista di attuali articoli, libri, tesi, atti di convegni e altre fonti scientifiche attinenti al tema "Principe de minimisation du risque empirique".
Accanto a ogni fonte nell'elenco di riferimenti c'è un pulsante "Aggiungi alla bibliografia". Premilo e genereremo automaticamente la citazione bibliografica dell'opera scelta nello stile citazionale di cui hai bisogno: APA, MLA, Harvard, Chicago, Vancouver ecc.
Puoi anche scaricare il testo completo della pubblicazione scientifica nel formato .pdf e leggere online l'abstract (il sommario) dell'opera se è presente nei metadati.
Articoli di riviste sul tema "Principe de minimisation du risque empirique"
Rasulo, Margherita, e Florian Pietron. "Dilemmes éthiques en terrains militants radicaux". L'Année sociologique Vol. 74, n. 2 (23 settembre 2024): 381–413. http://dx.doi.org/10.3917/anso.242.0381.
Testo completoNederlandt, Olivia, e Aurore Vanliefde. "La (non-)mixité entre hommes et femmes détenues dans les prisons belges. Une analyse des enjeux de genre dans les discours des personnes détenues et du personnel pénitentiaire". Droit et société N° 116, n. 1 (6 maggio 2024): 71–90. http://dx.doi.org/10.3917/drs1.116.0071.
Testo completoPasseron, Jean-Claude. "Ce que dit un tableau et ce qu’on en dit". Cambouis, la revue des sciences sociales aux mains sales, 9 gennaio 2021. http://dx.doi.org/10.52983/crev.vi0.35.
Testo completoMilani, Carlos R. S. "APRENDENDO COM A HISTÓRIA: críticas à experiência da Cooperação Norte-Sul e atuais desafios à Cooperação Sul-Sul". Caderno CRH 25, n. 65 (20 novembre 2012). http://dx.doi.org/10.9771/ccrh.v25i65.19347.
Testo completoTesi sul tema "Principe de minimisation du risque empirique"
Wade, Modou. "Apprentissage profond pour les processus faiblement dépendants". Electronic Thesis or Diss., CY Cergy Paris Université, 2024. http://www.theses.fr/2024CYUN1299.
Testo completoThis thesis focuses on deep learning for weakly dependent processes. We consider a class of deep neural network estimators with sparsity regularisation and/or penalty regularisation.Chapter1 is a summary of the work. It presents the deep learning framework and reviews the main results obtained in chapters 2, 3, 4, 5 and 6.Chapter 2 considers deep learning for psi-weakly dependent processes. We have established the convergence rate of the empirical risk minimization (ERM) algorithm on the class of deep neural network (DNN) estimators. For these estimators, we have provided a generalization bound and an asymptotic learning rate of order O(n^{-1/alpha}) for all alpha > 2 is obtained. A bound of the excess risk for a large class of target predictors is also established. Chapter 3 presents the sparse-penalized deep neural networks estimator under weak dependence. We consider nonparametric regression and classification problems for weakly dependent processes. We use a method of regularization by penalization. For nonparametric regression and binary classification, we establish an oracle inequality for the excess risk of the sparse-penalized deep neural networks (SPDNN) estimator. We have also provided a convergence rate for these estimators.Chapter 4 focuses on the penalized deep neural networks estimator with a general loss function under weak dependence. We consider the psi-weak dependence structure and, in the specific case where the observations are bounded, we deal with the theta_{infty}-weak dependence. For learning psi and theta_{infty}-weakly dependent processes, we have established an oracle inequality for the excess risks of the sparse-penalized deep neural networks estimator. We have shown that when the target function is sufficiently smooth, the convergence rate of these excess risks is close to O(n^{-1/3}).Chapter 5 presents robust deep learning from weakly dependent data. We assume that the output variable has finite r moments, with r >= 1. For learning strong mixing and psi-weakly dependent processes, a non-asymptotic bound for the expected excess risk of the deep neural networks estimator is established. We have shown that when the target function belongs to the class of H"older smooth functions, the convergence rate of the expected excess risk for exponentially strongly mixing data is close to or equal to that obtained with an independent and identically distributed sample. Chapter 6 focuses on deep learning for strongly mixing observation with sparse-penalized regularization and minimax optimality. We have provided an oracle inequality and a bound on the class of H"older smooth functions for the expected excess risk of the deep neural network estimator. We have also considered the problem of nonparametric regression from strongly mixing data with sub-exponential noise. When the target function belongs to the class of H"older composition functions, we have established an upper bound for the oracle inequality of the L_2 error. In the specific case of autoregressive regression with standard Laplace or normal error, we have provided a lower bound for the L_2 error in this classe, which matches up to a logarithmic factor the upper bound; thus the deep neural network estimator achieves optimal convergence rate
Papa, Guillaume. "Méthode d'échantillonnage appliqué à la minimisation du risque empirique". Electronic Thesis or Diss., Paris, ENST, 2018. http://www.theses.fr/2018ENST0005.
Testo completoIn this manuscript, we present and study applied sampling strategies, with problems related to statistical learning. The goal is to deal with the problems that usually arise in a context of large data when the number of observations and their dimensionality constrain the learning process. We therefore propose to address this problem using two sampling strategies: - Accelerate the learning process by sampling the most helpful. - Simplify the problem by discarding some observations to reduce complexity and the size of the problem. We first consider the context of the binary classification, when the observations used to form a classifier come from a sampling / survey scheme and present a complex dependency structure. for which we establish bounds of generalization. Then we study the implementation problem of stochastic gradient descent when observations are drawn non uniformly. We conclude this thesis by studying the problem of graph reconstruction for which we establish new theoretical results
Yu, Jiaqian. "Minimisation du risque empirique avec des fonctions de perte nonmodulaires". Thesis, Université Paris-Saclay (ComUE), 2017. http://www.theses.fr/2017SACLC012/document.
Testo completoThis thesis addresses the problem of learning with non-modular losses. In a prediction problem where multiple outputs are predicted simultaneously, viewing the outcome as a joint set prediction is essential so as to better incorporate real-world circumstances. In empirical risk minimization, we aim at minimizing an empirical sum over losses incurred on the finite sample with some loss function that penalizes on the prediction given the ground truth. In this thesis, we propose tractable and efficient methods for dealing with non-modular loss functions with correctness and scalability validated by empirical results. First, we present the hardness of incorporating supermodular loss functions into the inference term when they have different graphical structures. We then introduce an alternating direction method of multipliers (ADMM) based decomposition method for loss augmented inference, that only depends on two individual solvers for the loss function term and for the inference term as two independent subproblems. Second, we propose a novel surrogate loss function for submodular losses, the Lovász hinge, which leads to O(p log p) complexity with O(p) oracle accesses to the loss function to compute a subgradient or cutting-plane. Finally, we introduce a novel convex surrogate operator for general non-modular loss functions, which provides for the first time a tractable solution for loss functions that are neither supermodular nor submodular. This surrogate is based on a canonical submodular-supermodular decomposition
Zwald, Laurent. "Performances statistiques d'algorithmes d'apprentissage : "Kernel projection machine" et analyse en composantes principales à noyau". Paris 11, 2005. https://tel.archives-ouvertes.fr/tel-00012011.
Testo completoThis thesis takes place within the framework of statistical learning. It brings contributions to the machine learning community using modern statistical techniques based on progress in the study of empirical processes. The first part investigates the statistical properties of Kernel Principal Component Analysis (KPCA). The behavior of the reconstruction error is studied with a non-asymptotique point of view and concentration inequalities of the eigenvalues of the kernel matrix are provided. All these results correspond to fast convergence rates. Non-asymptotic results concerning the eigenspaces of KPCA themselves are also provided. A new algorithm of classification has been designed in the second part: the Kernel Projection Machine (KPM). It is inspired by the Support Vector Machines (SVM). Besides, it highlights that the selection of a vector space by a dimensionality reduction method such as KPCA regularizes suitably. The choice of the vector space involved in the KPM is guided by statistical studies of model selection using the penalized minimization of the empirical loss. This regularization procedure is intimately connected with the finite dimensional projections studied in the statistical work of Birge and Massart. The performances of KPM and SVM are then compared on some data sets. Each topic tackled in this thesis raises new questions
Rachdi, Nabil. "Apprentissage statistique et computer experiments : approche quantitative du risque et des incertitudes en modélisation". Toulouse 3, 2011. http://thesesups.ups-tlse.fr/1538/.
Testo completoThis thesis work consists in gathering statistical learning theory with the field of computer experiments. As the considered computer codes are only known through simulations, we propose an original statistical framework, including the classical ones, which takes into account the simulation aspect. We investigate learning algorithms for parameter estimation in computer codes which depend on both observed and simulation data. We validate these algorithms by proving excess risk bounds using concentration inequalities. We also study the duality between the estimation procedure and the wanted feature prediction. Here, we try to understand the impact of an estimation procedure on some given characteristic of the phenomenon of interest. Finally, the computation of optimal parameters in practice involves the minimization of a criterion which is generally highly non convex and with irregularities. We propose a stochastic algorithm which consists in combining regularization methods with a stochastic approximation method like the Kiefer-Wolfowitz one
Zwald, Laurent. "PERFORMANCES STATISTIQUES D'ALGORITHMES D'APPRENTISSAGE : ``KERNEL PROJECTION MACHINE'' ET ANALYSE EN COMPOSANTES PRINCIPALES A NOYAU". Phd thesis, Université Paris Sud - Paris XI, 2005. http://tel.archives-ouvertes.fr/tel-00012011.
Testo completodes contributions à la communauté du machine learning en utilisant des
techniques de statistiques modernes basées sur des avancées dans l'étude
des processus empiriques. Dans une première partie, les propriétés statistiques de
l'analyse en composantes principales à noyau (KPCA) sont explorées. Le
comportement de l'erreur de reconstruction est étudié avec un point de vue
non-asymptotique et des inégalités de concentration des valeurs propres de la matrice de
Gram sont données. Tous ces résultats impliquent des vitesses de
convergence rapides. Des propriétés
non-asymptotiques concernant les espaces propres de la KPCA eux-mêmes sont également
proposées. Dans une deuxième partie, un nouvel
algorithme de classification a été
conçu : la Kernel Projection Machine (KPM).
Tout en s'inspirant des Support Vector Machines (SVM), il met en lumière que la sélection d'un espace vectoriel par une méthode de
réduction de la dimension telle que la KPCA régularise
convenablement. Le choix de l'espace vectoriel utilisé par la KPM est guidé par des études statistiques de sélection de modéle par minimisation pénalisée de la perte empirique. Ce
principe de régularisation est étroitement relié à la projection fini-dimensionnelle étudiée dans les travaux statistiques de
Birgé et Massart. Les performances de la KPM et de la SVM sont ensuite comparées sur différents jeux de données. Chaque thème abordé dans cette thèse soulève de nouvelles questions d'ordre théorique et pratique.
Gazagnadou, Nidham. "Expected smoothness for stochastic variance-reduced methods and sketch-and-project methods for structured linear systems". Electronic Thesis or Diss., Institut polytechnique de Paris, 2021. http://www.theses.fr/2021IPPAT035.
Testo completoThe considerable increase in the number of data and features complicates the learning phase requiring the minimization of a loss function. Stochastic gradient descent (SGD) and variance reduction variants (SAGA, SVRG, MISO) are widely used to solve this problem. In practice, these methods are accelerated by computing these stochastic gradients on a "mini-batch": a small group of samples randomly drawn.Indeed, recent technological improvements allowing the parallelization of these calculations have generalized the use of mini-batches.In this thesis, we are interested in the study of variants of stochastic gradient algorithms with reduced variance by trying to find the optimal hyperparameters: step and mini-batch size. Our study allows us to give convergence results interpolating between stochastic methods drawing a single sample per iteration and the so-called "full-batch" gradient descent using all samples at each iteration. Our analysis is based on the expected smoothness constant which allows to capture the regularity of the random function whose gradient is calculated.We study another class of optimization algorithms: the "sketch-and-project" methods. These methods can also be applied as soon as the learning problem boils down to solving a linear system. This is the case of ridge regression. We analyze here variants of this method that use different strategies of momentum and acceleration. These methods also depend on the sketching strategy used to compress the information of the system to be solved at each iteration. Finally, we show that these methods can also be extended to numerical analysis problems. Indeed, the extension of sketch-and-project methods to Alternating-Direction Implicit (ADI) methods allows to apply them to large-scale problems, when the so-called "direct" solvers are too slow
Achab, Mastane. "Ranking and risk-aware reinforcement learning". Electronic Thesis or Diss., Institut polytechnique de Paris, 2020. http://www.theses.fr/2020IPPAT020.
Testo completoThis thesis divides into two parts: the first part is on ranking and the second on risk-aware reinforcement learning. While binary classification is the flagship application of empirical risk minimization (ERM), the main paradigm of machine learning, more challenging problems such as bipartite ranking can also be expressed through that setup. In bipartite ranking, the goal is to order, by means of scoring methods, all the elements of some feature space based on a training dataset composed of feature vectors with their binary labels. This thesis extends this setting to the continuous ranking problem, a variant where the labels are taking continuous values instead of being simply binary. The analysis of ranking data, initiated in the 18th century in the context of elections, has led to another ranking problem using ERM, namely ranking aggregation and more precisely the Kemeny's consensus approach. From a training dataset made of ranking data, such as permutations or pairwise comparisons, the goal is to find the single "median permutation" that best corresponds to a consensus order. We present a less drastic dimensionality reduction approach where a distribution on rankings is approximated by a simpler distribution, which is not necessarily reduced to a Dirac mass as in ranking aggregation.For that purpose, we rely on mathematical tools from the theory of optimal transport such as Wasserstein metrics. The second part of this thesis focuses on risk-aware versions of the stochastic multi-armed bandit problem and of reinforcement learning (RL), where an agent is interacting with a dynamic environment by taking actions and receiving rewards, the objective being to maximize the total payoff. In particular, a novel atomic distributional RL approach is provided: the distribution of the total payoff is approximated by particles that correspond to trimmed means
Neirac, Lucie. "Learning with a linear loss function : excess risk and estimation bounds for ERM and minimax MOM estimators, with applications". Electronic Thesis or Diss., Institut polytechnique de Paris, 2023. http://www.theses.fr/2023IPPAG012.
Testo completoCommunity detection, phase recovery, signed clustering, angular group synchronization, Maxcut, sparse PCA, the single index model, and the list goes on, are all classical topics within the field of machine learning and statistics. At first glance, they are pretty different problems with different types of data and different goals. However, the literature of recent years shows that they do have one thing in common: they all are amenable to Semi-Definite Programming (SDP). And because they are amenable to SDP, we can go further and recast them in the classical machine learning framework of risk minimization, and this with the simplest possible loss function: the linear loss function. This, in turn, opens up the opportunity to leverage the vast literature related to risk minimization to derive excess risk and estimation bounds as well as algorithms to unravel these problems. The aim of this work is to propose a unified methodology to obtain statistical properties of classical machine learning procedures based on the linear loss function, which corresponds, for example, to the case of SDP procedures that we look as ERM procedures. Embracing a machine learning view point allows us to go into greater depth and introduce other estimators which are effective in handling two key challenges within statistical learning: sparsity, and robustness to adversarial contamination and heavy-tailed data. We attack the structural learning problem by proposing a regularized version of the ERM estimator. We then turn to the robustness problem and introduce an estimator based on the median of means (MOM) principle, which we call the minmax MOM estimator. This latter estimator addresses the problem of robustness and can be constructed whatever the loss function, including the linear loss function. We also present a regularized version of the minmax MOM estimator. For each of those estimators we are able to provide excess risk and estimation bounds, which are derived from two key tools: local complexity fixed points and curvature equations of the excess risk function. To illustrate the relevance of our approach, we apply our methodology to five classical problems within the frame of statistical learning, for which we improve the state-of-the-art results