WorldCat Identities

École doctorale Sciences et technologies de l'information et mathématiques (Nantes)

Overview
Works: 498 works in 699 publications in 2 languages and 734 library holdings
Roles: Degree grantor, Other
Publication Timeline
.
Most widely held works by École doctorale Sciences et technologies de l'information et mathématiques (Nantes)
Sur les propriétés spectrales des opérateurs générés par un système d'équations différentielles by Fulya Seref( Book )

3 editions published in 2014 in English and held by 3 WorldCat member libraries worldwide

We consider non-self-adjoint operator Lm(Q) generated in Lm2 [0; 1] by the Sturm-Liouville equation with m _ m matrix potential and the boundary conditions, whose scalar case (m = 1) are strongly regular.First we obtain asymptotic formulas for the eigenvalues and eigenfunctions of Lm(Q) and then find a condition on the potential for which the root functions of the operator form a Riesz basis. We also study the approximation of eigenvalues of Lm(Q) by finite difference method
Méthode multi-échelle pour la résolution des équations de la cinétique neutronique by Steve Chauvet( Book )

3 editions published in 2008 in French and held by 3 WorldCat member libraries worldwide

In this PhD thesis and in order to improve the time/precision ratio of the numerical simulation calculations, we investigate multi-scale techniques for the resolution of the reactor kinetics equations.We choose to focus on the mixed dual diffusion approximation and the quasi-static methods. We introduce a space dependency for the amplitude function which only depends on the time variable in the standard quasi-static context. With this new factorization, we develop two mixed dual problems which can be solved with CEA's solver MINOS. An algorithm is implemented, performing the resolution of these problems defined on different scales (for time and space). We name this approach : the Local Quasi-Static method. We present here this new multi-scale approach and its implementation. The inherent details of amplitude and shape treatments are discussed and justified. Results and performances, compared to MINOS, are studied. They illustrate the improvement on the time/precision ratio for kinetics calculations. Furthermore, we open some new possibilities to parallelize computations with MINOS. For the future, we also introduce some improvement tracks with
Reconnaissance des entités nommées dans des documents multimodaux by Mohamed Hatmi( Book )

3 editions published in 2014 in French and held by 3 WorldCat member libraries worldwide

Named entity recognition is a subtask of information extraction. It consists of identifying some textual objects such as person, location and organization names. The work of this thesis focuses on the named entity recognition task for the oral modality. Some difficulties may arise for this task due to the intrinsic characteristics of speech processing (lack of capitalisation marks, lack of punctuation marks, presence of disfluences and of recognition errors...). In the first part, we study the characteristics of the named entity recognition downstream of the automatic speech recognition system. We present a methodology which allows named entity recognition following a hierarchical and compositional taxonomy. We measure the impact of the different phenomena specific to speech on the quality of named entity recognition. In the second part, we propose to study the tight pairing between the speech recognition task and the named entity recognition task. For that purpose, we take away the basic functionnalities of a speech recognition system to turn it into a named entity recognition system. Therefore, by mobilising the inherent knowledge of the speech processing to the named entity recognition task, we ensure a better synergy between the two tasks. We carry out different types of experiments to optimize and evaluate our approach
Étude d'opérateurs de traitement de signal appliquant l'échantillonnage aléatoire pour la radio logicielle by Najib Boumaaz( Book )

2 editions published in 2009 in French and held by 3 WorldCat member libraries worldwide

The presented work is the result of a cross between two recent fields: Software Radio (SR) and Random Sampling. The SR concept has emerged with the idea to allowing the realization of multi-mode and multi-standards infrastructure and terminals. At present, we consider the software defined radio architectures (SDR) which are more realistic. The development of systems operating with random sampling makes it possible to overcome the aliasing constraints encountered in the uniform sampling case. Indeed, these aliases are a significant obstacle for the choice and the sizing of the signal processing operators and their parameters. In order to optimize an SDR architecture, we tried to propose new and simple DFE structures applying random sampling. These structures should be simple, efficient and suitable for this kind of problems. In this thesis, we attempt to develop some signal processing operators predestined to be implemented in the DFE of an SDR architecture. We are particularly interested to multiband signal reconstruction algorithms and to channel filtering. We have studied some signal reconstruction methods treating finite sampling sequences. We have proposed a variety of DFE structures based on reconstruction signal algorithms or on adapted CIC filters. To complete the theoretical studies and simulations, we have analyzed software and hardware implementation aspects by developing some DFE operators on DSP card and FPGA
Le problème de Yamabe sur les espaces stratifiés by Ilaria Mondello( Book )

3 editions published in 2015 in English and held by 3 WorldCat member libraries worldwide

On étudie une classe d'espaces métriques singuliers, les espaces stratifiés, et on se propose d'étendre à ces derniers des résultats de géométrie riemannienne et d'analyse sur les variétés. Dans une première partie, on montre l'existence d'une borne inférieure pour le bas du spectre du Laplacien, sous une hypothèse géométrique de minoration de la courbure de Ricci. Cela permet également de démontrer l'existence d'une inégalité de Sobolev dont les constantes dépendent uniquement du volume et de la dimension de l'espace, et d'une borne supérieure pour le diamètre. En outre, on prouve que la borne pour le diamètre est atteinte si et seulement si celle pour le bas du spectre l'est aussi. La deuxième partie de ce manuscrit est dédiée aux conséquences des résultats précédents sur le problème de Yamabe pour un espace stratifié : ce problème consiste à chercher une métrique conforme à courbure scalaire constante, et l'existence d'une solution dépend d'un invariant conforme, la constante de Yamabe locale, dont la valeur est en général inconnue. On montre que celle-ci peut-être calculée en un grand nombre de cas, lorsque une hypothèse géométrique sur le lieu singulier est vérifiée. On utilise des techniques liées aux inégalités isopérimétrique et de Sobolev. Enfin, on donne une classe d'exemples pour lesquels on peut prouver qu'une métrique conforme à courbure scalaire constante existe
Modélisation dynamique et commande d'un propulseur naval by Matthieu Vonnet( Book )

2 editions published in 2008 in French and held by 3 WorldCat member libraries worldwide

Ce mémoire de thèse porte sur la modélisation dynamique et la commande en poussée des propulseurs navals. Un état de l'art de la modélisation des propulseurs électriques est d'abord présenté. Un modèle quasi-statique de propulseur est étudié et identifié expérimentalement en employant le bassin d'essais du Laboratoire IREENA. La nécessité d'un modèle dynamique d'hélice plus précis est mise en évidence. Deux approches sont alors envisagées. La première consiste à modéliser physiquement une hélice marine. Les théories des moments et la théorie des éléments de pale sont rappelées, puis la théorie dynamique du moment linéaire des hélices, développée dans un cadre rigoureux, est exposée. Un modèle dynamique d'hélice est déduit de cette dernière théorie et est étudié. La modélisation dynamique de l'hélice est ensuite envisagée par une approche fréquentielle, mieux adaptée à la commande. Des méthodes d'identification sans mesure hydrodynamique sont proposées et validées expérimentalement. La commande en poussée des propulseurs est abordée dans la dernière partie. Des méthodes de commande existantes sont brièvement étudiées et testées expérimentalement. Les tests réalisés mettent en évidence la nécessité de disposer d'une méthode de commande plus robuste vis-à-vis de la ventilation de l'hélice, qui occasionne des pertes de poussée importantes. Une nouvelle loi de commande, s'appuyant sur un observateur du couple hydrodynamique et sur le modèle dynamique de l'hélice, est finalement proposée. Des résultats expérimentaux, attestant la précision de la commande proposée, sont présentés
Optimisation du marquage des graphes d'événements values discrets : application aux systèmes flexibles de production by Laïla Toursi( Book )

2 editions published in 2009 in French and held by 3 WorldCat member libraries worldwide

L'un des problèmes importants qui se posent lors de la conception des systèmes flexibles de production est la minimisation des en-cours. Le nombre d'en-cours reflète la productivité du système puisqu'un nombre insuffisant d'en-cours ne permet pas d'atteindre la performance optimale. Cependant, un nombre trop important d'en-cours augmente les coûts de production car ces pièces utilisent des stocks ou des supports de transport. Ce problème a pour objectif la minimisation des coûts de production tout en gardant la même productivité. Les systèmes de production flexibles peuvent être modélisés par les graphes d'événements. Dans ce travail, nous nous intéressons au problème d'optimisation des ressources de production, notamment les moyens de transport, dans les systèmes de production à fabrication répétitive. En terme de graphes d'événements valués, ce problème est connu sous le nom d'optimisation du marquage et consiste à minimiser une somme pondérée de marquages des places (i.e. un critère linéaire P-invariant) sous la contrainte d'obtenir un temps de cycle donné. Nous proposons deux méthodes de résolution pour ce problème ainsi que quelques propriétés de la solution optimale. La première est une méthode de résolution par séparation et évaluation. Dans certains cas de GdEVs de structures particulières, cette méthode permet d'obtenir la solution optimale du problème. La deuxième méthode est une métaheuristique inspirée de l'algorithme du recuit simulé. Cette méthode est efficace pour les cas de GdEVs de tailles importantes. Les deux méthodes sont ensuite adaptées et appliquées à des systèmes de production flexibles
Commande non linéaire sans capteur de la machine asynchrone by Dramane Traoré( )

2 editions published in 2008 in French and held by 3 WorldCat member libraries worldwide

Existence locale et effet régularisant précisés pour des équations de type Schrödinger by Pierre-Yves Bienaimé( Book )

3 editions published in 2014 in French and held by 3 WorldCat member libraries worldwide

Dans cette thèse, on considère le problème de Cauchy dans les espaces de Sobolev habituels et dans des espaces de Sobolev à poids pour des équations non linéaires de la forme [Formule non transposable] : Ces équations sont de la forme des équations de Schrödinger. Nous étudions l'existence locale et l'effet régularisant vérifié par les solutions, pour cela nous suivons une méthode employée par C. E. Kenig, G. Ponce et L. Vega, et nous généralisons et précisons certains de leurs résultats. La non linéarité est un fonction régulière nulle à l'ordre 2 en 0 et l'opérateur [Formule non transposable] : Cet opérateur généralise le Laplacien mais n'est plus elliptique. Dans le cas où F est nulle à l'ordre 3 en 0, nous prouvons l'existence locale, l'unicité ainsi qu'un effet régularisant pour une donnée initiale dans [Formule non transposable] : Dans le cas où F est nulle à l'ordre 2 en 0, nous prouvons le même résultat mais pour une donnée initiale dans des espaces de Sobolev à poids. Le plan de démonstration reprend celui de C.E. KENIG, G. PO?CE et L. VEGA
Débruitage et correction d'images IRM : application à la caractérisation de produits agroalimentaires by Guylaine Collewet( )

2 editions published in 2008 in French and held by 3 WorldCat member libraries worldwide

Magnetic resonance imaging (MRI) is a non-invasive modality designed for clinical diagnosis. Other domains also exploit this technique, such as food products analysis. The applicative aim of our work is the study of the repartition of fat tissues in fish. We are particularly interested in the quantification of the tissues. An accurate quantification requires the denoising of the images and the correction of the intensity inhomogeneities due to the spatial variation of the radiofrequency magnetic field (RF). We use T1-weighted images. In this case, the effects of the RF inhomogeneities are complex since the bias that is induced in the images depends on the tissue. The proposed method takes place in the inverse problem framework where denoising and correcting are tackled jointly. It is based on a physical model of the MRI signal and a model of the sample considered as made of a finite number of tissues. The method relies on the minimisation of a penalised criterion. This criterion consists of a data-fitting term added with regularisation terms in order to ensure spatially smooth solutions while preserving the edges in the image. The method needs several images acquired with different protocols. The minimisation is based on a block-coordinate descent approach where each block consists in iterations of the conjugate gradient algorithm. Results obtained on images of fish validate our approach. We also present preliminary results on the optimisation of the choice of the protocols which lead to the best estimation of the variables. These results rely on the theory of experiment planning
Contribution à la considération explicite des plates-formes d'exécution logicielles lors d'un processus de déploiement d'application by Matthias Brun( )

1 edition published in 2010 in French and held by 3 WorldCat member libraries worldwide

Real-time and embedded software concerns, such as time and resource usage, are closely related to the execution platform. But the need for reuse leads to design platformindependent software. In order to ease platform integration of the application software, model-driven engeneering (MDE) aims at specializing application model using model transformations. But few works specify how to consider a platform within a transformation. On the one hand, this study compares several processes that transform a platform-independent model of an application into a platform-dependent model. This comparative work focuses on implicit or explicit consideration of real-time and embedded operating systems. On the other hand, this study explores a modeling method where platforms are explicitly described. A domain-specific modeling language (DSML) dedicated to the description of execution platforms is proposed. Then an application deployment process based on this language is introduced and evaluated. The process uses generic or generated transformation rules and a DSML dedicated to the description of application deployment
Contributions en classification automatique : agrégation bayésienne de mélanges de lois et visualisation interactive by Pierrick Bruneau( Book )

2 editions published in 2010 in French and held by 3 WorldCat member libraries worldwide

The internet and recent architectures such as sensor networks are currently witnessing tremendous and continuously growing amounts of data, often distributed on large scales. Combined with user expectations with respect to tooling, this encourages developing adequate techniques for analyzing and indexing. Classication and clustering tasks are about characterizing classes within data collections. These are often used as building blocks for designing tools aimed at making data accessible to users. In this document, we describe our contributions to mixture models aggregation. These models are classically used for content categorization. Using variational Bayesian principles, we aimed at designing low computation and transmission costs algorithms. Doing so, we aimed at proposing a building block for distributed density model estimation. We also contributed to visual classication applied to data streams. To this purpose, we employed bio-mimetic principles, and results from graph theory. More specically, visual and dynamic abstractions of an underlying clustering process were proposed. We strived to provide users with ecient interfaces, while allowing using their actions as a feedback
Modélisation et commande de l'infection par le VIH : aide au diagnostic et à la décision des traitements by Marie-José Mhawej( )

2 editions published in 2009 in French and held by 3 WorldCat member libraries worldwide

Unrelèvement d'une structure d'algèbre de Batalin-Vilkovisky sur la double construction cobar by Alexandre Quesney( Book )

3 editions published in 2014 in French and held by 3 WorldCat member libraries worldwide

In a first part we establish structural results on the cobar construction. The goal is to obtain a homotopy BV-algebra structure on the double cobar construction. In summary we have a criterion for obtaining of a homotopy BV-algebra (à la Gerstenhaber-Voronov) on the double cobar construction W2C of homotopy G-coalgebra C. This involves the structural co-operations of the homotopy G-coalgebra C. In a second part, we apply the previous criterion to the homotopy G-coalgebra C (X). The homotopy G-coalgebra structure on the simplicial chain complex C (X) is such that the resulting double cobar construction W2C (X) is a model for the double loop space W2jXj. Next, we give comparison results between the BV-algebra structure obtained on W2C (X) when X is a double suspension and the BV-algebra structure on H (W2jXj) given by the diagonal action of the circle. Finally, when Q is the coefficient ring, we deform the Hopf dg-algebra structure on the Baues cobar construction WC (X) into a involutive Hopf dg-algebra structure (r0 , S0). Then we obtain a homotopy BV-algebra structure on the double cobar construction W(WC (X),r0 , S0) for any simplicial set X
Méthodes numériques pour les écoulements en milieu poreux : estimations a posteriori et stratégie d'adaptation by Vincent Baron( Book )

3 editions published in 2015 in French and held by 3 WorldCat member libraries worldwide

This work is devoted to improving the efficiency of the numerical resolution of the Richards equation, which is a nonlinear parabolic equation used for the simulation of subsurface flows. The first part is dedicaced to the derivation of a DDFV scheme, which is valid on general meshes, coupled with the BDF2 formula used for the discretization of the instationary term. Numerical tests confirm the high-order accuracy of the method, as well as its stability in a variety of configurations. In the second part, we derive a posteriori estimates using the equilibrated fluxes method. A guaranteed upper bound is achieved, involving space-time estimators. This bound relies on an equilibrated fluxes relation built from relevant approximations of the continuous fluxes, reconstructed from the approximate solution. We present how to perform such reconstructions adapted to our DDFV-BDF2 discretization. They involve a correction term specific to the DDFV scheme used, as well as a rephrasing of the variable BDF2 formula under the form of a one-step scheme. Lastly, we apply the aformentioned estimate in some numerical tests by developing an adaptive algorithm, which, on a fixed mesh, aims to equilibrate the various error sources by adaptating the linearization stopping criterium and the time step. We analyze the influence of the parameters of this algorithm, and observe the number of linearization iterations and CPU time gained as compared to a classical simulation
Opérateur de Gauss-Bonnet semi-Fredholm et propriétés spectrales sur les graphes infinis by Hèla Ayadi( Book )

3 editions published in 2015 in French and held by 3 WorldCat member libraries worldwide

In the context of an infinite locally finite weighted graph, we are interested in the study of discrete Gauss-Bonnet operator which is a Dirac type operator ( its square is the Laplacian operator ). In particular, we are focused on the conditions to have semi-Fredholmness operator needed to approach the Hodge decomposition theorem, which is important for solving problems such that Kirchhoff's problem. In fact, we present a discrete version of the work of Gilles Carron which defines a new concept non-parabolicity at infinity to have the Gauss-Bonnet operator with closed range. Another part of this thesis consist to study the spectral properties of the Laplacian operator. We define two Laplacians one as an operator acting on functions on vertices and the other one acting on functions on edges. So, it is a natural question to characterize the relation between their spectrum in terms of a certain geometric property of the graph and properties of the operators. In fact, we show that the nonzero spectrum of the two laplacians are the same, by using Weyl criterion. In addition, we give an extension of the work of John Lott such that with suitable weight conditions, we prove that the spectral value 0 in the spectrum of one of these two Laplacians
Traitement automatique des termes composés : segmentation, traduction et variation by Elizaveta Loginova Clouet( Book )

3 editions published in 2014 in French and held by 3 WorldCat member libraries worldwide

The number of specialized terms continuously grows in the documents, at a pace which is difficult to follow for the terminology standardization organizations. The methods of bilingual term lexicon construction from the text corpora provide solutions. Our thesis falls into this topic: bilingual lexicon acquisition from comparable corpora. Compound terms (terms containing several roots, but a single graphical unit) are challenging for natural language processing applications. Given their graphical form, they are often handled in the same manner as single word terms, which prevents from apprehending their semantic complexity. Our involvement in an automatical terminology extraction evaluation allowed us to check our hypothesis: compound terms need a particular processing in a multilingual context. We proposed a method for compound terms recognition and splitting, which combines language-independent and language-specific features. It allowed us to obtain results comparable with those of state-of-the-art methods, while validating on a sample of languages from several families (germanic, slavic, romance languages), and adapting the method to specialized domains (tested on two domains: wind energy and breast cancer). We used the produced segmentations for compositional translation of compound terms, and for their multi-word variant recognition in the specialized texts. These two experiments illustrate that compound splitting is beneficial for the bilingual term lexicon acquisition task
Polymères dirigés et réseaux conducteurs de chaleur : systèmes de mécanique statistique à l'équilibre et hors équilibre by Alain Camanes( Book )

3 editions published in 2008 in French and held by 3 WorldCat member libraries worldwide

Génération de mouvements optimaux de marche pour des robots bipèdes 3D by David Tlalolini Romero( )

2 editions published in 2008 in French and held by 3 WorldCat member libraries worldwide

This work is devoted to the optimal motions generation of cycle walk for the anthropomorphic bipedal robots walking in the space. These robots are assumed to consist of a kinematics structure which retains only basic human locomotion mobility's on the level of the hip-knee-ankle-foot kinematics chain. The decomposition of the human gaits cycle, robot modelling and the constraints linked to the contact between the foot and the ground are presented. To dene dynamic model and impact model, the method of Newton-Euler is used. The motions generation of walking is posed under the form of a constrained parameter optimization problem. The resolution of this problem is obtained by Sequential Quadratic Programming (SQP) methods. The criterion is optimized in order to increase the autonomy of energy of the biped robot. To achieve and improve the convergence of the optimization algorithm, the gradient analytic calculation of the criteria was performed. In a preliminary study limited of the sagittal plane, dierent gaits are compared. These are dened by instantaneous double support phases or nite time double support phases and the single support phases including or not an under-actuated sub-phase in which stance foot rotates about the stance toe. These results are then generalized to the case of three-dimensional walk. The comparison of two gaits which includes the impacts and the single support phases with or without rolling of the foot is achieved. One can saw from these dierent studies that the introduction of phase of foot rolling in a single support phase allows much higher walking velocity with less energy consumption
De la perception locale des distorsions de codage à l'appréciation globale de la qualité visuelle des images et vidéos : apport de l'attention visuelle dans le jugement de qualité by Alexandre Ninassi( Book )

2 editions published in 2009 in French and held by 3 WorldCat member libraries worldwide

Cette étude traite de l'évaluation locale des distorsions perceptuelles, de l'évaluation globale de la qualité visuelle, et de l'influence de l'attention visuelle en évaluation de qualité.Afin d'évaluer localement les distorsions dans les images, nous avons simplifié un modèle existant du système visuel humain en utilisant la transformée en ondelettes et nous avons proposé une meilleure modélisation des effets de masquage par la prise en compte du masquage semi-local. A partir de ces modèles, nous avons conçu et validé des métriques de qualité d'images. Pour les vidéos, nous avons conçu une méthode d'évaluation locale des distorsions temporelles reposant sur un cumul temporel court terme des distorsions spatiales. Celui-ci simule l'évaluation des distorsions via des mécanismes de sélection de l'attention visuelle. Une métrique de qualité s'appuyant sur cette méthode a été conçue et validée. Celle-ci est basée sur un cumul temporel long terme incorporant un comportement asymétrique et un effet de saturation perceptuelle. L'influence de l'attention visuelle sur l'évaluation de la qualité a été analysée à partir des données issues de tests oculométriques réalisés sur des images et sur des vidéos, en exploration libre et en tâche de qualité. Les résultats ont confirmé, entre autres, l'influence de la tâche de qualité sur le déploiement de l'attention visuelle. L'impact de l'attention visuelle sur l'évaluation objective de la qualité a également été étudié en utilisant l'information de saillance réelle. Nous avons montré qu'une simple pondération linéaire des distorsions par l'attention visuelle ne permettait pas d'améliorer clairement les performances des métriques de qualité
 
moreShow More Titles
fewerShow Fewer Titles
Audience Level
0
Audience Level
1
  General Special  
Audience level: 0.00 (from 0.00 for Modélisat ... to 0.00 for Modélisat ...)

Alternative Names
École doctorale 503

École doctorale S.T.I.M.

École doctorale sciences et technologies de l'information et de mathématiques

École doctorale STIM

ED 503

ED STIM

ED503

Languages
French (43)

English (6)