Tesis sobre el tema "Réseaux de neurones bayésiens"
Crea una cita precisa en los estilos APA, MLA, Chicago, Harvard y otros
Consulte los 50 mejores tesis para su investigación sobre el tema "Réseaux de neurones bayésiens".
Junto a cada fuente en la lista de referencias hay un botón "Agregar a la bibliografía". Pulsa este botón, y generaremos automáticamente la referencia bibliográfica para la obra elegida en el estilo de cita que necesites: APA, MLA, Harvard, Vancouver, Chicago, etc.
También puede descargar el texto completo de la publicación académica en formato pdf y leer en línea su resumen siempre que esté disponible en los metadatos.
Explore tesis sobre una amplia variedad de disciplinas y organice su bibliografía correctamente.
Abid, Ilyes. "Modélisation de la prévision de défaillance des entreprises par des approches statiques et dynamiques : réseaux de neurones, réseaux bayésiens, modèles de durée et dichotomiques". Thesis, Paris 10, 2011. http://www.theses.fr/2011PA100151.
Texto completoThe objective of this thesis is to study bankruptcy prediction models from both static and dynamic viewpoints. More precisely, in the static approach, we used the methods of selecting discriminating variables using the neural networks. We thus proposed two new procedures relating to these methods. The first one is based on the criterion HVS called HVS-AUC and allowed to 1) build a more parsimonious model compared to the LDA, 2) identify a set of variables both static and non-cyclical with a strong explanatory power. Conversely, the second technique is based on the forward procedure, more precisely on forward-AUC. This method shows results comparable to the LDA but with fewer variables. It allows the detection of ratios considered as the most relevant according to LDA and HVS-AUC. We have also used methods of structure learning of Bayesian networks to improve the performance of classification of firms. We have mobilized a technique called "Max-Min Hill-Climbing" or MMHC. Specifically, we plan to analyze the performance of classification of an algorithm that mixes both MMHC and the canonical model of a naive Bayes network (NB). This new method could be called NB-MMHC (naive Bayes augmented by MMH C). The results confirm the prevailing view: as for the discriminatory power, no structure seems to be able to significantly compete with NB. In the second dynamic approach, we put more emphasis on factors not measurable a priori and also on explanatory factors impossible to capture within a static framework. In the first phase, we used the macroeconomic variables to better estimate the risk of default. In the second part, we used an alternative model to better estimate the shocks that firms could undergo over time. We therefore evaluate the propagation effects of theses shocks
Bourgeois, Yoann. "Les réseaux de neurones artificiels pour mesurer les risques économiques et financiers". Paris, EHESS, 2003. http://www.theses.fr/2003EHES0118.
Texto completoThe objective of this thesis is to provide complete methodologies to solve prediction and classification problems in economy and finance by using Artificial Neural networks. The plan of work shows that the thesisplays a great part in establishing in several ways a statistic methodology for neural networks. We proceed in four chapters. The first chapter describes supervised and unsupervised neural network methodology to modelize quantitative or qualitative variables. In the second chapter, we are interested by the bayesian approach for supervised neural networks and the developpement of a set of misspecification statistic tests for binary choice models. In chapter three, we show that multivariate supervised neural networks enable to take into account structural changes and the neural networks methodology is able to estimate some probabilities of exchange crisis. In chapter four, we develope a complete based neural network-GARCH model to manage a stocks portfolio. We introduce some terms as conditional returns or conditional risk for a stock or a portfolio. Next, we apply bayesian Self-Organizing Map in order to estimate the univariate probability density function of the DM/USD exchange rate
Mallet, Grégory. "Méthodes statistiques pour la prédiction de température dans les composants hyperfréquences". Phd thesis, INSA de Rouen, 2010. http://tel.archives-ouvertes.fr/tel-00586089.
Texto completoLiu, Haoran. "Statistical and intelligent methods for default diagnosis and loacalization in a continuous tubular reactor". Phd thesis, INSA de Rouen, 2009. http://tel.archives-ouvertes.fr/tel-00560886.
Texto completoTchoumatchenko, Irina. "Extraction des règles logiques dans des réseaux de neurones formels : application a la prédiction de la structure secondaire des protéines". Paris 6, 1994. http://www.theses.fr/1994PA066448.
Texto completoKozyrskiy, Bogdan. "Exploring the Intersection of Bayesian Deep Learning and Gaussian Processes". Electronic Thesis or Diss., Sorbonne université, 2023. https://accesdistant.sorbonne-universite.fr/login?url=https://theses-intra.sorbonne-universite.fr/2023SORUS064archi.pdf.
Texto completoDeep learning played a significant role in establishing machine learning as a must-have instrument in multiple areas. The use of deep learning poses several challenges. Deep learning requires a lot of computational power for training and applying models. Another problem with deep learning is its inability to estimate the uncertainty of the predictions, which creates obstacles in risk-sensitive applications. This thesis presents four projects to address these problems: We propose an approach making use of Optical Processing Units to reduce energy consumption and speed up the inference of deep models. We address the problem of uncertainty estimates for classification with Bayesian inference. We introduce techniques for deep models that decreases the cost of Bayesian inference. We developed a novel framework to accelerate Gaussian Process regression. We propose a technique to impose meaningful functional priors for deep models through Gaussian Processes
Kacimi, Sahra. "Contribution à la restitution des précipitations tropicales par radiométrie micro-ondes : préparation à la mission Megha-Tropiques". Versailles-St Quentin en Yvelines, 2012. http://www.theses.fr/2012VERS0064.
Texto completoWithin the framework of the global warming, the analysis of water and energy budget is of major importance. Considering the Megha-Tropiques (MT) mission whose one of the scientific objectives is to improve the knowledge of water and energy cycle in the intertropical region, the estimation of instantaneous surface rainfall is of the great importance. My PhD work focuses on the optimization of a multi-region, the estimation of instantaneous surface rainfall is of great importance. My PhD work focuses on the optimization of a multi-plateform Bayesian retrieval algorithm called BRAIN (Bayesian Retrieval Algorithm Including Neural Networks) (Viltard et al. , 2006) used for MT. This algorithm uses passive microwave data from satellites such as TRMM, SSM/I and AQUA. It uses a Bayesian Monte Carlo approach to retrieve several atmospheric parameters such as the instantaneous rainfall rate. In order to get a more accurate rainfall restitution, two research axes were investigated : the detection of a priori rainy areas that takes place before the rainfall estimation itself, and the impact of the database and inversion parameters. First, the database on which the algorithm relies needs to be more representative especially as far as high rain rates are concerned. To improve the representativeness of the inversion database, we need first to eliminate repetitive profiles, that is to say extract prototypes from it. To be made, we use Self Organizing Maps SO%s developed by T. Kohonen (2001). Second, the improvement of the rainy-non-rainy pixels classification before the inversion was made using neural networks
Tran, Gia-Lac. "Advances in Deep Gaussian Processes : calibration and sparsification". Electronic Thesis or Diss., Sorbonne université, 2020. https://accesdistant.sorbonne-universite.fr/login?url=https://theses-intra.sorbonne-universite.fr/2020SORUS410.pdf.
Texto completoGaussian Processes (GPs) are an attractive specific way of doing non-parametric Bayesian modeling in a supervised learning problem. It is well-known that GPs are able to make inferences as well as predictive uncertainties with a firm mathematical background. However, GPs are often unfavorable by the practitioners due to their kernel's expressiveness and the computational requirements. Integration of (convolutional) neural networks and GPs are a promising solution to enhance the representational power. As our first contribution, we empirically show that these combinations are miscalibrated, which leads to over-confident predictions. We also propose a novel well-calibrated solution to merge neural structures and GPs by using random features and variational inference techniques. In addition, these frameworks can be intuitively extended to reduce the computational cost by using structural random features. In terms of computational cost, the exact Gaussian Processes require the cubic complexity to training size. Inducing point-based Gaussian Processes are a common choice to mitigate the bottleneck by selecting a small set of active points through a global distillation from available observations. However, the general case remains elusive and it is still possible that the required number of active points may exceed a certain computational budget. In our second study, we propose Sparse-within-Sparse Gaussian Processes which enable the approximation with a large number of inducing points without suffering a prohibitive computational cost
Wolinski, Pierre. "Structural Learning of Neural Networks". Thesis, université Paris-Saclay, 2020. http://www.theses.fr/2020UPASS026.
Texto completoThe structure of a neural network determines to a large extent its cost of training and use, as well as its ability to learn. These two aspects are usually in competition: the larger a neural network is, the better it will perform the task assigned to it, but the more it will require memory and computing time resources for training. Automating the search of efficient network structures -of reasonable size and performing well- is then a very studied question in this area. Within this context, neural networks with various structures are trained, which requires a new set of training hyperparameters for each new structure tested. The aim of the thesis is to address different aspects of this problem. The first contribution is a training method that operates within a large perimeter of network structures and tasks, without needing to adjust the learning rate. The second contribution is a network training and pruning technique, designed to be insensitive to the initial width of the network. The last contribution is mainly a theorem that makes possible to translate an empirical training penalty into a Bayesian prior, theoretically well founded. This work results from a search for properties that theoretically must be verified by training and pruning algorithms to be valid over a wide range of neural networks and objectives
Fond, Antoine. "Localisation par l'image en milieu urbain : application à la réalité augmentée". Thesis, Université de Lorraine, 2018. http://www.theses.fr/2018LORR0028/document.
Texto completoThis thesis addresses the problem of localization in urban areas. Inferring accurate positioning in the city is important in many applications such as augmented reality or mobile robotics. However, systems based on inertial sensors (IMUs) are subject to significant drifts and GPS data can suffer from a valley effect that limits their accuracy. A natural solution is to rely on the camera pose estimation in computer vision. We notice that buildings are the main visual landmarks of human beings but also objects of interest for augmented reality applications. We therefore aim to compute the camera pose relatively to a database of known reference buildings from a single image. The problem is twofold : find the visible references in the current image (place recognition) and compute the camera pose relatively to them. Conventional approaches to these two sub-problems are challenged in urban environments due to strong perspective effects, frequent repetitions and visual similarity between facades. While specific approaches to these environments have been developed that exploit the high structural regularity of such environments, they still suffer from a number of limitations in terms of detection and recognition of facades as well as pose computation through model registration. The original method developed in this thesis is part of these specific approaches and aims to overcome these limitations in terms of effectiveness and robustness to clutter and changes of viewpoints and illumination. For do so, the main idea is to take advantage of recent advances in deep learning by convolutional neural networks to extract high-level information on which geometric models can be based. Our approach is thus mixed Bottom- Up/Top-Down and is divided into three key stages. We first propose a method to estimate the rotation of the camera pose. The 3 main vanishing points of the image of urban environnement, known as Manhattan vanishing points, are detected by a convolutional neural network (CNN) that estimates both these vanishing points and the image segmentation relative to them. A second refinement step uses this information and image segmentation in a Bayesian model to estimate these points effectively and more accurately. By estimating the camera’s rotation, the images can be rectified and thus free from perspective effects to find the translation. In a second contribution, we aim to detect the facades in these rectified images to recognize them among a database of known buildings and estimate a rough translation. For the sake of efficiency, a series of cues based on facade specific characteristics (repetitions, symmetry, semantics) have been proposed to enable the fast selection of facade proposals. Then they are classified as facade or non-facade according to a new contextual CNN descriptor. Finally, the matching of the detected facades to the references is done by a nearest neighbor search using a metric learned on these descriptors. Eventually we propose a method to refine the estimation of the translation relying on the semantic segmentation inferred by a CNN for its robustness to changes of illumination ans small deformations. If we can already estimate a rough translation from these detected facades, we choose to refine this result by relying on the se- mantic segmentation of the image inferred from a CNN for its robustness to changes of illuminations and small deformations. Since the facade is identified in the previous step, we adopt a model-based approach by registration. Since the problems of registration and segmentation are linked, a Bayesian model is proposed which enables both problems to be jointly solved. This joint processing improves the results of registration and segmentation while remaining efficient in terms of computation time. These three parts have been validated on consistent community data sets. The results show that our approach is fast and more robust to changes in shooting conditions than previous methods
Rio, Maxime. "Modèles bayésiens pour la détection de synchronisations au sein de signaux électro-corticaux". Phd thesis, Université de Lorraine, 2013. http://tel.archives-ouvertes.fr/tel-00859307.
Texto completoJauffret, Adrien. "De l'auto-évaluation aux émotions : approche neuromimétique et bayésienne de l'apprentissage de comportements complexes impliquant des informations multimodales". Thesis, Paris 11, 2014. http://www.theses.fr/2014PA112120/document.
Texto completoThe goal of this thesis is to build a bio-inspired architecture allowing a robot to autonomouslynavigate over large distances. In a cognitive science point of view, the model also aim at improv-ing the understanding of the underlying biological mechanisms. Previous works showed thata computational model of hippocampal place cells, based on neurobiological studies made onrodent, allows a robot to learn robust navigation behaviors. The robot can learn a round or ahoming behavior from a few associations between places and actions. The learning and recog-nition of a place were only defined by visual information and shows limitations for navigatinglarge environments.Adding other sensorial modalities is an effective solution for improving the robustness of placesrecognition in complex environments. This solution led us to the elementary blocks requiredwhen trying to perform multimodal information merging. Such merging has been done, first,by a simple conditioning between 2 modalities and next improved by a more generic model ofinter-modal prediction. In this model, each modality learns to predict the others in usual situa-tions, in order to be able to detect abnormal situations and to compensate missing informationof the others. Such a low level mechanism allows to keep a coherent perception even if onemodality is wrong. Moreover, the model can detect unexpected situations and thus exhibit someself-assessment capabilities: the assessment of its own perception. Following this model of self-assessment, we focus on the fundamental properties of a system for evaluating its behaviors.The first fundamental property that pops out is the statement that evaluating a behavior is anability to recognize a dynamics between sensations and actions, rather than recognizing a sim-ple sensorial pattern. A first step was thus to take into account the sensation/action couplingand build an internal minimalist model of the interaction between the agent and its environment.Such of model defines the basis on which the system will build predictions and expectations.The second fundamental property of self-assessment is the ability to extract relevant informa-tion by the use of statistical processes to perform predictions. We show how a neural networkcan estimate probability density functions through a simple conditioning rule. This probabilis-tic learning allows to achieve bayesian inferences since the system estimates the probability ofobserving a particular behavior from statistical information it recognizes about this behavior.The robot estimates the different statistical momentums (mean, variance, skewness, etc...) of abehavior dynamics by cascading few simple conditioning. Then, the non-recognition of such adynamics is interpreted as an abnormal behavior.But detecting an abnormal behavior is not sufficient to conclude to its inefficiency. The systemmust also monitor the temporal evolution of such an abnormality to judge the relevance of thebehavior. We show how an emotional meta-controller can use this novelty detection to regu-late behaviors and so select the best appropriate strategy in a given context. Finally, we showhow a simple frustration mechanism allows the robot to call for help when it detects potentialdeadlocks. Such a mechanism highlights situations where a skills improvement is possible, soas some developmental processes
Drouet, Isabelle. "Causalité et probabilités : réseaux bayésiens, propensionnisme". Phd thesis, Université Panthéon-Sorbonne - Paris I, 2007. http://tel.archives-ouvertes.fr/tel-00265287.
Texto completoTsala, Éric. "Désambiguïsation sémantique et réseaux bayésiens dynamiques". Mémoire, Université de Sherbrooke, 2009. http://savoirs.usherbrooke.ca/handle/11143/4833.
Texto completoBen, Mrad Ali. "Observations probabilistes dans les réseaux bayésiens". Thesis, Valenciennes, 2015. http://www.theses.fr/2015VALE0018/document.
Texto completoIn a Bayesian network, evidence on a variable usually signifies that this variable is instantiated, meaning that the observer can affirm with certainty that the variable is in the signaled state. This thesis focuses on other types of evidence, often called uncertain evidence, which cannot be represented by the simple assignment of the variables. This thesis clarifies and studies different concepts of uncertain evidence in a Bayesian network and offers various applications of uncertain evidence in Bayesian networks.Firstly, we present a review of uncertain evidence in Bayesian networks in terms of terminology, definition, specification and propagation. It shows that the vocabulary is not clear and that some terms are used to represent different concepts.We identify three types of uncertain evidence in Bayesian networks and we propose the followingterminology: likelihood evidence, fixed probabilistic evidence and not-fixed probabilistic evidence. We define them and describe updating algorithms for the propagation of uncertain evidence. Finally, we propose several examples of the use of fixed probabilistic evidence in Bayesian networks. The first example concerns evidence on a subpopulation applied in the context of a geographical information system. The second example is an organization of agent encapsulated Bayesian networks that have to collaborate together to solve a problem. The third example concerns the transformation of evidence on continuous variables into fixed probabilistic evidence. The algorithm BN-IPFP-1 has been implemented and used on medical data from CHU Habib Bourguiba in Sfax
Touya, Thierry. "Méthodes d'optimisation pour l'espace et l'environnement". Phd thesis, Université Paul Sabatier - Toulouse III, 2008. http://tel.archives-ouvertes.fr/tel-00366141.
Texto completoLa première traite d'une antenne spatiale réseau active.
Il faut d'abord calculer les lois d'alimentation pour satisfaire les contraintes de rayonnement. Nous transformons un problème avec de nombreux minima locaux en un problème d'optimisation convexe, dont l'optimum est le minimum global du problème initial, en utilisant le principe de conservation de l'énergie.
Nous résolvons ensuite un problème d'optimisation topologique: il faut réduire le nombre d'éléments rayonnants (ER). Nous appliquons une décomposition en valeurs singulières à l'ensemble des modules optimaux relaxés, puis un algorithme de type gradient topologique décide les regroupements entre ER élémentaires.
La deuxième partie porte sur une simulation type boîte noire d'un accident chimique.
Nous effectuons une étude de fiabilité et de sensibilité suivant un grand nombre de paramètres (probabilités de défaillance, point de conception, et paramètres influents). Sans disposer du gradient, nous utilisons un modèle réduit.
Dans un premier cas test nous avons comparé les réseaux neuronaux et la méthode d'interpolation sur grille éparse Sparse Grid (SG). Les SG sont une technique émergente: grâce à leur caractère hiérarchique et un algorithme adaptatif, elles deviennent particulièrement efficaces pour les problèmes réels (peu de variables influentes).
Elles sont appliquées à un cas test en plus grande dimension avec des améliorations spécifiques (approximations successives et seuillage des données).
Dans les deux cas, les algorithmes ont donné lieu à des logiciels opérationnels.
Perez, Joan. "Spatial structures in India in the age of globalisation : a data-driven approach". Thesis, Avignon, 2015. http://www.theses.fr/2015AVIG1151/document.
Texto completoCountries that have experienced a delayed entry within the world economy have usually sustained an enhanced and faster globalisation process. This is the case for BRIC countries which are, compared to other emerging countries, organised on large economies and thus provide a stronger potential market. From this perspective, India appears to be the perfect case study with an economic growth expected to overcome China’s growth in the near future. However, the «clichés» are persistent within a country mostly depicted as bipolar. On the one hand, it is considered as a new eldorado, the «Shining India», a place where multinationals aim to implement themselves due to the substantial increase of the consumer market. At the same time, India is also characterised by overcrowding, the major presence of slum areas and mass poverty, both in urban and rural areas. It is indeed possible that some areas will accommodate a bigger and bigger share of the growing middle class, while others will accentuate economic and social inequalities. Yet, can these extremes be truly representative of the diversity of such a large country? In fact, in some urban oriented spaces, the evolution of the tertiary sector is not strong enough to maintain a high level of employment while in rural spaces; an intensive farming model contributes to gradually reducing the number of labourers and landowners. As a result, the increase of the standard of living related to both economic and demographic growth is not homogeneously distributed over a territory where socio-economic divisions are already made worse by a tight caste system. With evidence dating back to 2400 BCE, it must be remembered that India is a country of old urbanisation. This has given rise to a rich and complex history and India is now home to a variety of languages, religions, castes, communities, tribes, traditions, urbanisation patterns and, more recently, globalisation-related dynamics. Perhaps no other country in the world seems to be characterised by such a great diversity. This begs the following questions: how is it possible to quantify and visualise the spatial gap of such a complex and subcontinent sized country? What are the main drivers affecting this spatial gap? It would indeed be simplistic to study India only through macro-economic indicators such as GDP. To deal with this complexity, a conceptualisation has been performed to strictly select 55 criteria that can affect the transformation sustained by the Indian territory in this enhanced age of globalisation. These selected factors have fed a multi-critera database characterised by aspects coming from economy, geography, sociology, culture etc. at the district scale level (640 spatial units) and on a ten year timeframe (2001-2011). The assumption is as follows: each Indian district can be driven by different factors. The human capacity to understand a complex issue has been reached here since we cannot take into account and at the same time the behaviour of a large number of elements influencing one another. AI Based Algorithm methods (Bayesian and Neural Networks) have thus been resorted to as a good alternative to process a large number of factors. In order to be as accurate as possible and to keep a transversal point of view, the methodology is divided into a robust procedure including fieldwork steps. The results of the models show that the 55 factors interact, bringing the emergence of unobservable factors representative of broader concepts, which find consistency only in the case of India. It also shows that the Indian territory can be segmented into a multitude of sub-spaces. Some of these profiles are close to the caricatured India. However, in most cases, results show a heterogeneous country with sub-spaces possessing a logic of their own and far away from any cliché
Trinh, Quoc Anh. "Méthodes neuronales dans l'analyse de survie". Evry, Institut national des télécommunications, 2007. http://www.theses.fr/2007TELE0004.
Texto completoThis thesis proposes a generalization of the conventional survival models where the linear prdictive variables are replaced by nonlinear multi-layer perceptions of variables. This modelling by neural networks predict the survival times with talking into account the time effects and the interactions between variables. The neural network models will be validated by cross validation technique or the bayesian slection criterion based on the model's posteriori probability. The prediction is refined by a boostrap aggregating (Bagging) and bayesian models average to increase the precision. Moreower, the censoring, the particularity of the survival analysis, needs a survival model which could take into account all available knowledges on the data for estimation to obtain a better prediction. The bayesian approach is thus a proposed approach because it allows a better generalization of the neural networks because of the avoidance of the overlifting. Moreover, the hierarchical models in bayesian learning of the neural networks is appropriate perfectly for a selection of relevant variables which gives a better explanation of the times effects and the interactions between variables
Prestat, Emmanuel. "Les réseaux bayésiens : classification et recherche de réseaux locaux en cancérologie". Phd thesis, Université Claude Bernard - Lyon I, 2010. http://tel.archives-ouvertes.fr/tel-00707732.
Texto completoLeray, Ph. "Réseaux bayésiens : Apprentissage et diagnostic de systemes complexes". Habilitation à diriger des recherches, Université de Rouen, 2006. http://tel.archives-ouvertes.fr/tel-00485862.
Texto completoSanchez-Soto, Eduardo. "Réseaux Bayésiens Dynamiques pour la Vérification du Locuteur". Phd thesis, Télécom ParisTech, 2005. http://tel.archives-ouvertes.fr/tel-00011440.
Texto completoSmail, Linda. "Algorithmique pour les Réseaux Bayésiens et leurs extensions". Phd thesis, Université de Marne la Vallée, 2004. http://tel.archives-ouvertes.fr/tel-00007170.
Texto completoLe chapitre 1 présente la théorie des réseaux bayésiens. Nous introduisons une nouvelle notion, celle de réseau bayésien de niveau deux, utile pour l'introduction de notre algorithme de calcul sur les réseaux bayésiens ; nous donnons également quelques résultats fondamentaux et nous situons dans notre formalisme un exemple d'école de réseau bayésien dit «Visite en Asie» .
Dans le second chapitre, nous exposons une propriété graphique appelée «d-séparation» grâce à laquelle on peut déterminer, pour tout couple de variables aléatoires ou de groupes de variables, et tout ensemble de conditionnement, s'il y a nécessairement, ou non, indépendance conditionnelle. Nous présentons également dans ce chapitre des résultats concernant le calcul de probabilités ou probabilités conditionnelles dans les réseaux bayésiens en utilisant les propriétés de la d-séparation. Ces résultats, qui concernent des écritures à notre connaissance originales de la factorisation de la loi jointe et de la loi conditionnée d'une famille de variables aléatoires du réseau bayésien (en liaison avec la notion de réseau bayésien de niveau deux) doivent trouver leur utilité pour les réseaux bayésiens de grande taille.
Le troisième chapitre donne la présentation détaillée et la justification d'un des algorithmes connus de calcul dans les réseaux bayésiens : il s'agit de l'algorithme LS (Lauritzen and Spigelhalter), basé sur la méthode de l'arbre de jonction. Pour notre part, après avoir présenté la notion de suite recouvrante propre possédant la propriété d'intersection courante, nous proposons un algorithme en deux versions (dont l'une est originale) qui permet de construire une suite de parties d'un réseau bayésien possédant cette propriété. Cette présentation est accompagnée d'exemples.
Dans le chapitre 4, nous donnons une présentation détaillée de l'algorithme des restrictions successives que nous proposons pour le calcul de lois (dans sa première version), et de lois conditionnelles (dans sa deuxième version). Cela est présenté après l'introduction d'une nouvelle notion : il s'agit de la descendance proche. Nous présentons également une application de l'algorithme des restrictions successives sur l'exemple «Visite en Asie» présenté en chapitre 1, et nous comparons le nombre d'opérations élémentaires effectuées avec celui qui intervient dans l'application de l'algorithme LS sur le même exemple. Le gain de calcul qui, à la faveur de cet exemple, apparaît au profit de l'algorithme des restrictions successives, sera comme toujours, d'autant plus marqué que la taille des réseaux et le nombre de valeurs prises par les variables seront plus élevés. C'est ce qui justifie l'insertion de notre algorithme au seins de « ProBT » , un logiciel d'inférence probabiliste, réalisé et diffusé par l'équipe Laplace localisée dans le laboratoire Gravir à INRIA Rhône Alpes.
En annexes nous rappelons les propriétés des graphes orientés sans circuits, les notions de base sur l'indépendance conditionnelle et l'équivalence de plusieurs définitions des réseaux bayésiens.
Sánchez-Soto, Eduardo. "Réseaux bayésiens dynamiques pour la vérification du locuteur". Paris, ENST, 2005. http://www.theses.fr/2005ENST0032.
Texto completoThis thesis is concerned with the statistical modeling of speech signal applied to Speaker Verification (SV) using Bayesian Networks (BNs). The main idea of this work is to use BNs as a mathematical tool to model pertinent speech features keeping its relations. It combines theoretical and experimental work. The difference between systems and humans performance in SV is the quantity of information and the relationships between the sources of information used to make decisions. A single statistical framework that keeps the conditional dependence and independence relations between those variables is difficult to attain. Therefore, the use of BNs as a tool for modeling the available information and their independence and dependence relationships is proposed. The first part of this work reviews the main modules of a SV system, the possible sources of information as well as the basic concepts of graphical models. The second part deals with Modeling. A new approach to the problems associated with the SV systems is proposed. The problem of inference and learning (parameters and structure)in BNs are presented. In order to obtain an adapted structure the relations of conditional independence among the variables are learned directly from the data. These relations are then used in order to build an adapted BN. In particular, a new model adaptation technique for BN has been proposed. This adaptation is based on a measure between Conditional Probability Distributions for discrete variables and on Regression Matrix for continuous variables used to model the relationships. In a large database for the SV task, the results have confirmed the potential of use the BNs approach
Wenzek, Didier. "Construction de réseaux de neurones". Phd thesis, Grenoble INPG, 1993. http://tel.archives-ouvertes.fr/tel-00343569.
Texto completoHallouli, Khalid. "Reconnaissance de caractères par méthodes markoviennes et réseaux bayésiens". Phd thesis, Télécom ParisTech, 2004. http://pastel.archives-ouvertes.fr/pastel-00000740.
Texto completoVerron, Sylvain. "Diagnostic et surveillance des processus complexes par réseaux bayésiens". Phd thesis, Université d'Angers, 2007. http://tel.archives-ouvertes.fr/tel-00517101.
Texto completoRahier, Thibaud. "Réseaux Bayésiens pour fusion de données statiques et temporelles". Thesis, Université Grenoble Alpes (ComUE), 2018. http://www.theses.fr/2018GREAM083/document.
Texto completoPrediction and inference on temporal data is very frequently performed using timeseries data alone. We believe that these tasks could benefit from leveraging the contextual metadata associated to timeseries - such as location, type, etc. Conversely, tasks involving prediction and inference on metadata could benefit from information held within timeseries. However, there exists no standard way of jointly modeling both timeseries data and descriptive metadata. Moreover, metadata frequently contains highly correlated or redundant information, and may contain errors and missing values.We first consider the problem of learning the inherent probabilistic graphical structure of metadata as a Bayesian Network. This has two main benefits: (i) once structured as a graphical model, metadata is easier to use in order to improve tasks on temporal data and (ii) the learned model enables inference tasks on metadata alone, such as missing data imputation. However, Bayesian network structure learning is a tremendous mathematical challenge, that involves a NP-Hard optimization problem. We present a tailor-made structure learning algorithm, inspired from novel theoretical results, that exploits (quasi)-determinist dependencies that are typically present in descriptive metadata. This algorithm is tested on numerous benchmark datasets and some industrial metadatasets containing deterministic relationships. In both cases it proved to be significantly faster than state of the art, and even found more performant structures on industrial data. Moreover, learned Bayesian networks are consistently sparser and therefore more readable.We then focus on designing a model that includes both static (meta)data and dynamic data. Taking inspiration from state of the art probabilistic graphical models for temporal data (Dynamic Bayesian Networks) and from our previously described approach for metadata modeling, we present a general methodology to jointly model metadata and temporal data as a hybrid static-dynamic Bayesian network. We propose two main algorithms associated to this representation: (i) a learning algorithm, which while being optimized for industrial data, is still generalizable to any task of static and dynamic data fusion, and (ii) an inference algorithm, enabling both usual tasks on temporal or static data alone, and tasks using the two types of data.%We then provide results on diverse cross-field applications such as forecasting, metadata replenishment from timeseries and alarms dependency analysis using data from some of Schneider Electric’s challenging use-cases.Finally, we discuss some of the notions introduced during the thesis, including ways to measure the generalization performance of a Bayesian network by a score inspired from the cross-validation procedure from supervised machine learning. We also propose various extensions to the algorithms and theoretical results presented in the previous chapters, and formulate some research perspectives
Nguyen, Hoai-Tuong. "Réseaux bayésiens et apprentissage ensembliste pour l'étude différentielle de réseaux de régulation génétique". Phd thesis, Université de Nantes, 2012. http://tel.archives-ouvertes.fr/tel-00675310.
Texto completoNguyen, Hoai Tuong. "Réseaux bayésiens et apprentissage ensembliste pour l’étude différentielle de réseaux de régulation génétique". Nantes, 2012. http://www.theses.fr/2012NANT2004.
Texto completoLn the recent years, the Bayesian networks (SN) have become one of the most powerful machine learning methods to modeling graphically and probabilistically different kinds of complex systems. One of the common issues in BN structure learning is the small-data problem. Ln fact, the result of learning is sensible to sample size of dataset. Ln machine learning, the set-based learning methods such as Bootstrap ou genetic algorithms are the often used methods to dealing with the small-data problem. However, the existing methods limit generally to the fusion of a set of models, but do not allow to compare two set of models. Inspired from the obtained results of the set-based methods, we proposed a novel method based on the quasi-essential graph (QEG) and the usage of the multiple testing in order to compare two sets of BN. QEG allows to resume and visualize graphically a set of BN. The multiple testing allows to verify if the differences between two set of BN are statistically significative and to determine the position of the differences. The application on the synthetic and experimental data demonstrated the different interests of proposed method in gene regulatory networks reconstruction and perspectively in the other applications with the small dataset
Tsopze, Norbert. "Treillis de Galois et réseaux de neurones : une approche constructive d'architecture des réseaux de neurones". Thesis, Artois, 2010. http://www.theses.fr/2010ARTO0407/document.
Texto completoThe artificial neural networks are successfully applied in many applications. But theusers are confronted with two problems : defining the architecture of the neural network able tosolve their problems and interpreting the network result. Many research works propose some solutionsabout these problems : to find out the architecture of the network, some authors proposeto use the problem domain theory and deduct the network architecture and some others proposeto dynamically add neurons in the existing networks until satisfaction. For the interpretabilityproblem, solutions consist to extract rules which describe the network behaviour after training.The contributions of this thesis concern these problems. The thesis are limited to the use of theartificial neural networks in solving the classification problem.In this thesis, we present a state of art of the existing methods of finding the neural networkarchitecture : we present a theoritical and experimental study of these methods. From this study,we observe some limits : difficulty to use some method when the knowledges are not available ;and the network is seem as ’black box’ when using other methods. We a new method calledCLANN (Concept Lattice-based Artificial Neural Network) which builds from the training dataa semi concepts lattice and translates this semi lattice into the network architecture. As CLANNis limited to the two classes problems, we propose MCLANN which extends CLANN to manyclasses problems.A new method of rules extraction called ’MaxSubsets Approach’ is also presented in thisthesis. Its particularity is the possibility of extracting the two kind of rules (If then and M-of-N)from an internal structure.We describe how to explain the MCLANN built network result aboutsome inputs
Maalej, Mohamed-Amine y Véronique Delcroix. "Diagnostic multiple des systèmes complexes à base de réseaux bayésiens". Valenciennes, 2006. http://ged.univ-valenciennes.fr/nuxeo/site/esupversions/46028b3e-dbca-41c6-a69f-b740ee4d0519.
Texto completoModel Based Diagnosis approach revolutionizes the field of the diagnosis as overcoming the lack of knowledge by using Model. Our research focuses on the task of multiple diagnosis, from failure observations, for complex and highly-reliable large systems. We take the advantages of the Bayesian networks models to improve the diagnosis of this type of systems. These models integrate the components failures prior probabilities, and allow estimating posterior probabilities of diagnoses, by an approached calculation. We present a methodology of diagnosis using Bayesian network. Our approach integrates a model design method, in addition to two diagnosis algorithms : the first algorithm allows calculating the most probable diagnoses for a failing system; the second provides quickly representatives of the diagnosis classes, it reveals also the cases while additional observations are necessary. Finally we test these algorithms in terms of computing time and results quality for digital circuits of various sizes
Wuillemin, Pierre-Henri. "Améliorations et implémentations d'algorithmes de propagation dans les réseaux bayésiens". Paris 6, 2000. http://www.theses.fr/2000PA066574.
Texto completoDelaplace, Alain. "Approche évolutionnaire de l'apprentissage de structure pour les réseaux bayésiens". Tours, 2007. http://www.theses.fr/2007TOUR4022.
Texto completoIn this thesis, we propose a study of the problem of learning the structure of a bayesian network through the use of evolutionary methods. We first designed a genetic algorithm to search the space of structures before establishing various strategies aiming at improving the performances of this algorithm. We consequently developed a search strategy aiming at exploiting the properties of the space of completed partially oriented graphs using a sequential niching principle which we later hybridized with an island model scheme. Another method defines a distribution probability over the mutation operations which are applied to the individuals and that is a function of the qualitative results of previously applied operations
Voegtlin, Thomas. "Réseaux de neurones et auto-référence". Lyon 2, 2002. http://theses.univ-lyon2.fr/documents/lyon2/2002/voegtlin_t.
Texto completoThe purpose of this thesis is to present a class of unsupervised learning algorithms for recurrent networks. In the first part (chapters 1 to 4), I propose a new approach to this question, based on a simple principle: self-reference. A self-referent algorithm is not based on the minimization of an objective criterion, such as an error function, but on a subjective function, that depends on what the network has previously learned. An example of a supervised recurrent network where learning is self-referent is the Simple Recurrent Network (SRN) by Elman (1990). In the SRN, self-reference is applied to the supervised error back-propagation algorithm. In this aspect, the SRN differs from other generalizations of back-propagation to recurrent networks, that use an objective criterion, such as Back-Propagation Through Time, or Real-Time Recurrent Learning. In this thesis, I show that self-reference can be combined with several well-known unsupervised learning methods: the Self-Organizing Map (SOM), Principal Components Analysis (PCA), and Independent Components Analysis (ICA). These techniques are classically used to represent static data. Self-reference allows one to generalize these techniques to time series, and to define unsupervised learning algorithms for recurrent networks
Teytaud, Olivier. "Apprentissage, réseaux de neurones et applications". Lyon 2, 2001. http://theses.univ-lyon2.fr/documents/lyon2/2001/teytaud_o.
Texto completoCôté, Marc-Alexandre. "Réseaux de neurones génératifs avec structure". Thèse, Université de Sherbrooke, 2017. http://hdl.handle.net/11143/10489.
Texto completoSaad, Ali. "Detection of Freezing of Gait in Parkinson's disease". Thesis, Le Havre, 2016. http://www.theses.fr/2016LEHA0029/document.
Texto completoFreezing of Gait (FoG) is an episodic phenomenon that is a common symptom of Parkinson's disease (PD). This research is headed toward implementing a detection, diagnosis and correction system that prevents FoG episodes using a multi-sensor device. This particular study aims to detect/diagnose FoG using different machine learning approaches. In this study we validate the choice of integrating multiple sensors to detect FoG with better performance. Our first level of contribution is introducing new types of sensors for the detection of FoG (telemeter and goniometer). An advantage in our work is that due to the inconsistency of FoG events, the extracted features from all sensors are combined using the Principal Component Analysis technique. The second level of contribution is implementing a new detection algorithm in the field of FoG detection, which is the Gaussian Neural Network algorithm. The third level of contribution is developing a probabilistic modeling approach based on Bayesian Belief Networks that is able to diagnosis the behavioral walking change of patients before, during and after a freezing event. Our final level of contribution is utilizing tree-structured Bayesian Networks to build a global model that links and diagnoses multiple Parkinson's disease symptoms such as FoG, handwriting, and speech. To achieve our goals, clinical data are acquired from patients diagnosed with PD. The acquired data are subjected to effective time and frequency feature extraction then introduced to the different detection/diagnosis approaches. The used detection methods are able to detect 100% of the present appearances of FoG episodes. The classification performances of our approaches are studied thoroughly and the accuracy of all methodologies is considered carefully and evaluated
Auliac, Cédric. "Approches évolutionnaires pour la reconstruction de réseaux de régulation génétique par apprentissage de réseaux bayésiens". Phd thesis, Université d'Evry-Val d'Essonne, 2008. http://tel.archives-ouvertes.fr/tel-00421388.
Texto completoJodouin, Jean-François. "Réseaux de neurones et traitement du langage naturel : étude des réseaux de neurones récurrents et de leurs représentations". Paris 11, 1993. http://www.theses.fr/1993PA112079.
Texto completoLasserre, Marvin. "Apprentissages dans les réseaux bayésiens à base de copules non-paramétriques". Electronic Thesis or Diss., Sorbonne université, 2022. http://www.theses.fr/2022SORUS029.
Texto completoModeling multivariate continuous distributions is a task of central interest in statistics and machine learning with many applications in science and engineering. However, high-dimensional distributions are difficult to handle and can lead to intractable computations. The Copula Bayesian Networks (CBNs) take advantage of both Bayesian networks (BNs) and copula theory to compactly represent such multivariate distributions. Bayesian networks rely on conditional independences in order to reduce the complexity of the problem, while copula functions allow to model the dependence relation between random variables. The goal of this thesis is to give a common framework to both domains and to propose new learning algorithms for copula Bayesian networks. To do so, we use the fact that CBNs have the same graphical language as BNs which allows us to adapt their learning methods to this model. Moreover, using the empirical Bernstein copula both to design conditional independence tests and to estimate copulas from data, we avoid making parametric assumptions, which gives greater generality to our methods
Brette, Romain. "Modèles Impulsionnels de Réseaux de Neurones Biologiques". Phd thesis, Université Pierre et Marie Curie - Paris VI, 2003. http://tel.archives-ouvertes.fr/tel-00005340.
Texto completoTardif, Patrice. "Autostructuration des réseaux de neurones avec retards". Thesis, Université Laval, 2007. http://www.theses.ulaval.ca/2007/24240/24240.pdf.
Texto completoMaktoobi, Sheler. "Couplage diffractif pour réseaux de neurones optiques". Thesis, Bourgogne Franche-Comté, 2020. http://www.theses.fr/2020UBFCD019.
Texto completoPhotonic networks with high performance can be considered as substrates for future computing systems. In comparison with electronics, photonic systems have substantial privileges, for instance the possibility of a fully parallel implementation of networks. Recently, neural networks have moved into the center of attention of the photonic community. One of the most important requirements for parallel large-scale photonic networks is to realize the connectivities. Diffraction is considered as a method to process the connections between the nodes (coupling) in optical neural networks. In the current thesis, we evaluate the scalability of a diffractive coupling in more details as follow:First, we begin with a general introductions for artificial intelligence, machine learning, artificial neural network and photonic neural networks. To establish a working neural network, learning rules are an essential part to optimize a configuration for obtaining a low error from the system, hence learning rules are introduced (Chapter 1). We investigate the fundamental concepts of diffractive coupling in our spatio-temporal reservoir. In that case, theory of diffraction is explained. We use an analytical scheme to provide the limits for the size of diffractive networks which is a part of our photonic neural network (Chapter 2). The concepts of diffractive coupling are investigated experimentally by two different experiments to confirm the analytical limits and to obtain maximum number of nodes which can be coupled in the photonic network (Chapter 3). Numerical simulations for such an experimental setup is modeled in two different schemes to obtain the maximum size of network numerically, which approaches a surface of 100 mm2 (Chapter 4). Finally, the complete photonic neural network is demonstrated. We design a spatially extended reservoir for 900 nodes. Consequently, our system generalizes the prediction for the chaotic Mackey–Glass sequence (Chapter 5)
Ouali, Jamel. "Architecture intégrée flexible pour réseaux de neurones". Grenoble INPG, 1991. http://www.theses.fr/1991INPG0035.
Texto completoCheung-Mon-Chan, Pascal. "Réseaux bayésiens et filtres particulaires pour l'égalisation adaptative et le décodage conjoints". Phd thesis, Télécom ParisTech, 2003. http://pastel.archives-ouvertes.fr/pastel-00000732.
Texto completoTian, Simiao. "Prédiction de la composition corporelle par modélisation locale et les réseaux bayésiens". Thesis, Paris, AgroParisTech, 2013. http://www.theses.fr/2013AGPT0068/document.
Texto completoThe assessment of human body composition is important for evaluating health and nutritional status. Among health issues, overweight and obesity are worldwide problems. Increased fat mass, especially in the trunk location, has been associated with an increased risk of metabolic diseases, such as type 2 diabetes and cardiovascular disease. The lean body mass, especially appendicular muscle mass, is also directly related to health and particularly with the mortality rate. Also, aging is associated with substantial changes in body composition. Reduction in body lean or body fat-free mass occurs during aging (Kyle et al., 2001) together with an increase of body fat related to accumulation of adipose tissues, particularly in abdominal region (Kuk et al., 2009); therefore assessing these changes in segmental body composition may be important because the study will lead to a pre-diagnosis for the prevention of morbidity and mortality risk. Accurate measurements of body composition can be obtained from different methods, such as underwater weighing and dual-energy X-ray absorptiometry (DXA). However, their applications are not always convenient, because they require fixed equipment and they are also time consuming and expensive. As a result, they are not convenient for use as a part of routine clinical examinations or population studies. The potential uses of statistical methods for body composition assessment have been highlighted (Snijder et al., 2006), and several attempts to predict body composition, particularly body fat percentage (BF%), have been made (Gallagher et al., 2000a; Jackson et al., 2002; Mioche et al., 2011b).The first aim in this thesis was to develop a multivariate model for predicting simultaneously body, trunk and appendicular fat and lean masses from easily measured anthropometric covariables. We proposed a linear solution published in the British Journal of Nutrition. There are two main advantages in our proposed multivariate approach. The first consists in using very simple covariables, such as body weight and height, because these measurements are easy and not expensive. The usefulness of waist circumference is also investigated and combined with age, height and weight as predictor variables. The second advantage is that the multivariate approach enables to take into account the correlation structure between the responses into account, which is useful for a number of inference tasks, e.g., to give simultaneous confidence regions for all the responses together. Then the prediction accuracy of the multivariate approach is justified by comparing with that of the available univariate models that predict body fat percentage (BF%). With a good accuracy, the multivariate outcomes might then be used in studies necessitating the assessment of metabolic risk factors in large populations.The second aim in this thesis was to study age-related changes in segmental body compositions, associated with anthropometric covariables. Two Bayesian modeling methods are proposed for the exploration of age-related changes. The main advantage of these methods is to propose a surrogate for a longitudinal analysis from the cross-sectional datasets. Moreover, the Bayesian modeling enables to provide a prediction distribution, rather than a simple estimate, this is more relevant for exploring the uncertainty or accuracy problems. Also we can incorporate the previous findings in the prior distribution, by combining it with the datasets, we could obtain more suitable conclusions.The previous predictions were based on models supposing any correlation structure within the variables, the third aim in this thesis was to propose a parsimonious sub-model of the multivariable model described by a Gaussian Bayesian network (GBN), more precisely Crossed Gaussian Bayesian Networks (CGBN). Last and final summary in the thesis
Bendou, Mohamed. "Extraction de connaissances à partir des données à l'aide des réseaux bayésiens". Paris 11, 2003. http://www.theses.fr/2003PA112053.
Texto completoThe main objective of this thesis basically focuses on developing a new kind of learning algorithms of Bayésiens networks, more accurate, efficient and robust in presence of the noise and, therefore, adapted to KDD tasks. Since most of local optima in the space of networks bayésiens structures are caused directly by the existence of equivalence classes (sets of structures encoding the same conditional independence relations, represented by the partially oriented graphs), we concentrated important part of our researches on the development of a new family of learning algorithms: EQ. These algorithms directly explore the space of equivalence classes. We also developed theoretical and algorithmic tools for the analysis and the treatment of partially oriented graphs. We could demonstrate that a meaningful precision gains brought by this kind of approach can be obtained in a comparable time than the classical approaches. We, thus, contributed to the present interest renewal for the learning of equivalence classes of bayesian networks (considered for a long time as too complex by the scientific community). Finally, another aspect of our research has been dedicated to the analysis of noise effects in data on the learning of the Bayesians networks. We analyzed and explained the increase of the complexity of learned Bayesian networks learned from noisy data and shown that, unlike classical over-fitting which affects other classes of learning methods, this phenomenon is theoretically justified by the alteration of the conditional independence relations between the variables and is beneficial for the predictive power of the learned models
Bigot, Pascal. "Utilisation des réseaux de neurones pour la télégestion des réseaux techniques urbains". Lyon 1, 1995. http://www.theses.fr/1995LYO10036.
Texto completoKoiran, Pascal. "Puissance de calcul des réseaux de neurones artificiels". Lyon 1, 1993. http://www.theses.fr/1993LYO19003.
Texto completoGraïne, Slimane. "Inférence grammaticale régulière par les réseaux de neurones". Paris 13, 1994. http://www.theses.fr/1994PA132020.
Texto completo