Dissertationen zum Thema „Apprentissage de structures causales“
Geben Sie eine Quelle nach APA, MLA, Chicago, Harvard und anderen Zitierweisen an
Machen Sie sich mit Top-50 Dissertationen für die Forschung zum Thema "Apprentissage de structures causales" bekannt.
Neben jedem Werk im Literaturverzeichnis ist die Option "Zur Bibliographie hinzufügen" verfügbar. Nutzen Sie sie, wird Ihre bibliographische Angabe des gewählten Werkes nach der nötigen Zitierweise (APA, MLA, Harvard, Chicago, Vancouver usw.) automatisch gestaltet.
Sie können auch den vollen Text der wissenschaftlichen Publikation im PDF-Format herunterladen und eine Online-Annotation der Arbeit lesen, wenn die relevanten Parameter in den Metadaten verfügbar sind.
Sehen Sie die Dissertationen für verschiedene Spezialgebieten durch und erstellen Sie Ihre Bibliographie auf korrekte Weise.
Assaad, Charles. „Découvertes de relations causales entre séries temporelles“. Electronic Thesis or Diss., Université Grenoble Alpes, 2021. http://www.theses.fr/2021GRALM019.
Der volle Inhalt der QuelleThis thesis aims to give a broad coverage of central concepts and principles of causation and in particular the ones involved in the emerging approaches to causal discovery from time series.After reviewing concepts and algorithms, we first present a new approach that infer a summary graph of the causal system underlying the observational time series while relaxing the idealized setting of equal sampling rates and discuss the assumptions underlying its validity. The gist of our proposal lies in the introduction of the causal temporal mutual information measure that can detect the independence and the conditional independence between two time series, and in making an apparent connection between entropy and the probability raising principle that can be used for building new rules for the orientation of the direction of causation. Moreover, through the development of this base method, we propose several extensions, namely to handle hidden confounders, to infer a window causal graph given a summary graph, and to consider sequences instead of time series.Secondly, we focus on the discovery of causal relations from a statistical distribution that is not entirely faithful to the real causal graph and on distinguishing a common cause from an intermediate cause even in the absence of a time indicator. The key aspect of our answer to this problem is the reliance on the additive noise principle to infer a directed supergraph that contains the causal graph. To converge toward the causal graph, we use in a second step a new measure called the temporal causation entropy that prunes for each node of the directed supergraph, the parents that are conditionally independent of their child. Furthermore, we explore complementary extensions of our second base method that involve a pairwise strategy which reduces through multitask learning and a denoising technique, the number of functions that need to be estimated. We perform an extensive experimental comparison of the proposed algorithms on both synthetic and real datasets and demonstrate their promising practical performance: gaining in time complexity while preserving accuracy
Li, Honghao. „Interpretable biological network reconstruction from observational data“. Electronic Thesis or Diss., Université Paris Cité, 2021. http://www.theses.fr/2021UNIP5207.
Der volle Inhalt der QuelleThis thesis is focused on constraint-based methods, one of the basic types of causal structure learning algorithm. We use PC algorithm as a representative, for which we propose a simple and general modification that is applicable to any PC-derived methods. The modification ensures that all separating sets used during the skeleton reconstruction step to remove edges between conditionally independent variables remain consistent with respect to the final graph. It consists in iterating the structure learning algorithm while restricting the search of separating sets to those that are consistent with respect to the graph obtained at the end of the previous iteration. The restriction can be achieved with limited computational complexity with the help of block-cut tree decomposition of the graph skeleton. The enforcement of separating set consistency is found to increase the recall of constraint-based methods at the cost of precision, while keeping similar or better overall performance. It also improves the interpretability and explainability of the obtained graphical model. We then introduce the recently developed constraint-based method MIIC, which adopts ideas from the maximum likelihood framework to improve the robustness and overall performance of the obtained graph. We discuss the characteristics and the limitations of MIIC, and propose several modifications that emphasize the interpretability of the obtained graph and the scalability of the algorithm. In particular, we implement the iterative approach to enforce separating set consistency, and opt for a conservative rule of orientation, and exploit the orientation probability feature of MIIC to extend the edge notation in the final graph to illustrate different causal implications. The MIIC algorithm is applied to a dataset of about 400 000 breast cancer records from the SEER database, as a large-scale real-life benchmark
Soldano, Henry. „Apprentissage : Paradigmes, Structures et abstractions“. Habilitation à diriger des recherches, Université Paris-Nord - Paris XIII, 2009. http://tel.archives-ouvertes.fr/tel-00514160.
Der volle Inhalt der QuelleTommasi, Marc. „Structures arborescentes et apprentissage automatique“. Habilitation à diriger des recherches, Université Charles de Gaulle - Lille III, 2006. http://tel.archives-ouvertes.fr/tel-00117063.
Der volle Inhalt der QuelleÀ la base de ce travail se trouve la question de l'accès et de la manipulation automatique d'informations au format XML au sein d'un réseau d'applications réparties dans internet. La réalisation de ces applications est toujours du ressort de programmeurs spécialistes d'XML et reste hors de portée de l'utilisateur final. De plus, les développements récents d'internet poursuivent l'objectif d'automatiser les communications entre applications s'échangeant des flux de données XML. Le recours à des techniques d'apprentissage automatique est une réponse possible à cette situation.
Nous considèrons que les informations sont décrites dans un langage XML, et dans la perspective de ce mémoire, embarquées dans des données structurées sous forme arborescente. Les applications sont basées alors sur des opérations élémentaires que sont l'interrogation ou les requêtes dans ces documents arborescents ou encore la transformation de tels documents.
Nous abordons alors la question sous l'angle de la réalisation automatique de programmes d'annotation d'arbres, permettant de dériver des procédures de transformation ou d'exécution de requêtes. Le mémoire décrit les contributions apportées pour la manipulation et l'apprentissage d'ensembles d'arbres d'arité non bornée (comme le sont les arbres XML), et l'annotation par des méthodes de classification supervisée ou d'inférence statistique.
Ben, Hassena Anouar. „Apprentissage par analogie de structures d'arbres“. Rennes 1, 2012. http://www.theses.fr/2011REN1E007.
Der volle Inhalt der QuelleAffeldt, Séverine. „Reconstruction de réseaux fonctionnels et analyse causale en biologie des systèmes“. Thesis, Paris 6, 2015. http://www.theses.fr/2015PA066171/document.
Der volle Inhalt der QuelleThe inference of causality is an everyday life question that spans a broad range of domains for which interventions or time-series acquisition may be impracticable if not unethical. Yet, elucidating causal relationships in real-life complex systems can be convoluted when relying solely on observational data. I report here a novel network reconstruction method, which combines constraint-based and Bayesian frameworks to reliably reconstruct networks despite inherent sampling noise in finite observational datasets. The approach is based on an information theory result tracing back the existence of colliders in graphical models to negative conditional 3-point information between observed variables. This enables to confidently ascertain structural independencies in causal graphs, based on the ranking of their most likely contributing nodes with (significantly) positive conditional 3-point information. Dispensible edges from a complete undirected graph are progressively pruned by iteratively taking off the most likely positive conditional 3-point information from the 2-point (mutual) information between each pair of nodes. The resulting skeleton is then partially directed by orienting and propagating edge directions based on the sign and magnitude of the conditional 3-point information of unshielded triples. This new approach outperforms constraint-based and Bayesian inference methods on a range of benchmark networks and provides promising predictions when applied to the reconstruction of complex biological systems, such as hematopoietic regulatory subnetworks, zebrafish neural networks, mutational pathways or the interplay of genomic properties on the evolution of vertebrates
Dinh, Quang-Thang. „Apprentissage statistique relationnel : apprentissage de structures de réseaux de Markov logiques“. Phd thesis, Université d'Orléans, 2011. http://tel.archives-ouvertes.fr/tel-00659738.
Der volle Inhalt der QuelleTillmann, Barbara. „Perception des structures musicales : apprentissage et modélisation“. Dijon, 1999. http://www.theses.fr/1999DIJOL026.
Der volle Inhalt der QuelleWestern listeners acquire a sensitivity to the regularities of the tonal system of Western music by mere exposure to tonal musical pieces. Once acquired, this knowledge allows the construction of a mental representation of the structure of tonal musical pieces. The present research examines the perception of large musical structures, the contribution of tonal knowledge in this processing and the acquisition of this knowledge. The first part of the research investigates the perception of global and local structures in real musical pieces. Different experimental paradigms (judgments of expressivity and coherence, resolution of “musical puzzles”, target detection and recognition) provide evidence that local structures prevail over global structures. Global structures seem to have relatively little importance for the listener in pieces of 3 min. Or 20 sec in length. The second part examines the contribution of tonal knowledge to the processing of event structures in short sequences (8 to 14 chords). The harmonic priming paradigm reveals an influence of global and local structures on chord processing. Simulations with a connectionist model of tonal knowledge representation (Bharucha, 1987) suggest that the influence of the structures emerges from simple activation of tonal knowledge and its accumulation over time – without hierarchical integration in an overall structure. Connectionist models allow modeling of the capacity of the cognitive system to extract underlying regularities from the environment. In the third part, learning simulations of the tonal system are realized with the help of an unsupervised learning algorithm. The network, combined with a spreading activation mechanism, allows modeling of experimental data on the processing of event structures in terms of information accumulation rather than in terms of a hierarchical integration strictly speaking
Déguernel, Ken. „Apprentissage de structures musicales en contexte d'improvisation“. Thesis, Université de Lorraine, 2018. http://www.theses.fr/2018LORR0011/document.
Der volle Inhalt der QuelleCurrent musical improvisation systems are able to generate unidimensional musical sequences by recombining their musical contents. However, considering several dimensions (melody, harmony...) and several temporal levels are difficult issues. In this thesis, we propose to combine probabilistic approaches with formal language theory in order to better assess the complexity of a musical discourse, both from a multidimensional and multi-level point of view in the context of improvisation where the amount of data is limited. First, we present a system able to follow the contextual logic of an improvisation modelled by a factor oracle whilst enriching its musical discourse with multidimensional knowledge represented by interpolated probabilistic models. Then, this work is extended to create another system using a belief propagation algorithm representing the interaction between several musicians, or between several dimensions, in order to generate multidimensional improvisations. Finally, we propose a system able to improvise on a temporal scenario with multi-level information modelled with a hierarchical grammar. We also propose a learning method for the automatic analysis of hierarchical temporal structures. Every system is evaluated by professional musicians and improvisers during listening sessions
Daoudi, Abderrazak. „Acquisition de contraintes par apprentissage de structures“. Thesis, Montpellier, 2016. http://www.theses.fr/2016MONTT316/document.
Der volle Inhalt der QuelleConstraint Programming is a general framework used to model and solve complex combinatorial problems.However, modeling a problem as a constraint network requires significant expertise in the field.Such level of expertise is a bottleneck to the broader uptake of the constraint technology.To alleviate this issue, several constraint acquisition systems have been proposed to assist thenon-expert user in the modeling task. Nevertheless, in these systems the user is only asked to answervery basic questions. The drawback is that when no background knowledge is provided,the user may need to answer a large number of such questions to learn all the constraints.In this thesis, we show that using the structure of the problem under consideration may improvethe acquisition process a lot. To this aim, we propose several techniques.Firstly, we introduce the concept of generalization query based on an aggregation of variables into types.Secondly, to deal with generalization queries, we propose a constraint generalization algorithm, named GENACQ, together with several strategies. Thirdly, to make the build of generalization queries totally independent of the user, we propose the algorithm MINE&ASK, which is able to learn the structure, during the constraint acquisition process, and to use the learned structure to generate generalization queries. Fourthly, toward a generic concept of query, we introduce the recommendation query based on the link prediction on the current constraint graph. Fifthly, we propose a constraint recommender algorithm, called PREDICT&ASK, that asks recommendation queries, each time the structure of the current graph has been modified. Finally, we incorporate all these new generic techniques into QUACQ algorithm leading to three boosted versions, G-QUACQ, M- QUACQ, and P-QUACQ. To evaluate all these techniques, we have made experiments on several benchmarks. The results show that the extended versions improve drastically the basic QUACQ
Mountassir, Mahjoub El. „Surveillance d'intégrité des structures par apprentissage statistique : application aux structures tubulaires“. Thesis, Université de Lorraine, 2019. http://www.theses.fr/2019LORR0047.
Der volle Inhalt der QuelleTo ensure better working conditions of civil and engineering structures, inspections must be made on a regular basis. However, these inspections could be labor-intensive and cost-consuming. In this context, structural health monitoring (SHM) systems using permanently attached transducers were proposed to ensure continuous damage diagnostic of these structures. In SHM, damage detection is generally based on comparison between the healthy state signals and the current signals. Nevertheless, the environmental and operational conditions will have an effect on the healthy state signals. If these effects are not taken into account they would result in false indication of damage (false alarm). In this thesis, classical machine learning methods used for damage detection have been applied in the case of pipelines. The effects of some measurements parameters on the robustness of these methods have been investigated. Afterthat, two approaches were proposed for damage diagnostic depending on the database of reference signals. If this database contains large variation of these EOCs, a sparse estimation of the current signal is calculated. Then, the estimation error is used as an indication of the presence of damage. Otherwise, if this database is acquired at limited range of EOCs, moving window PCA can be applied to update the model of the healthy state provided that the EOCs show slow and continuous variation. In both approaches, damage localization was ensured using a sliding window over the damaged pipe signal
Affeldt, Séverine. „Reconstruction de réseaux fonctionnels et analyse causale en biologie des systèmes“. Electronic Thesis or Diss., Paris 6, 2015. http://www.theses.fr/2015PA066171.
Der volle Inhalt der QuelleThe inference of causality is an everyday life question that spans a broad range of domains for which interventions or time-series acquisition may be impracticable if not unethical. Yet, elucidating causal relationships in real-life complex systems can be convoluted when relying solely on observational data. I report here a novel network reconstruction method, which combines constraint-based and Bayesian frameworks to reliably reconstruct networks despite inherent sampling noise in finite observational datasets. The approach is based on an information theory result tracing back the existence of colliders in graphical models to negative conditional 3-point information between observed variables. This enables to confidently ascertain structural independencies in causal graphs, based on the ranking of their most likely contributing nodes with (significantly) positive conditional 3-point information. Dispensible edges from a complete undirected graph are progressively pruned by iteratively taking off the most likely positive conditional 3-point information from the 2-point (mutual) information between each pair of nodes. The resulting skeleton is then partially directed by orienting and propagating edge directions based on the sign and magnitude of the conditional 3-point information of unshielded triples. This new approach outperforms constraint-based and Bayesian inference methods on a range of benchmark networks and provides promising predictions when applied to the reconstruction of complex biological systems, such as hematopoietic regulatory subnetworks, zebrafish neural networks, mutational pathways or the interplay of genomic properties on the evolution of vertebrates
Perreault, Samuel. „Structures de corrélation partiellement échangeables : inférence et apprentissage automatique“. Doctoral thesis, Université Laval, 2020. http://hdl.handle.net/20.500.11794/66443.
Der volle Inhalt der QuelleFrançois, Clément. „Apprentissage implicite des structures linguistiques et musicales : approche multi-méthodologique“. Thesis, Aix-Marseille 2, 2011. http://www.theses.fr/2011AIX20673/document.
Der volle Inhalt der QuelleThe aims of the present thesis were two-folded. Firstly, we wanted to compare behavioral and electrophysiological measures related to the implicit learning of linguistic and musical structures contained within an artificial sung language. While behavioral measures suggest that only the linguistic structure was learned, electrophysiological data revealed similar N400 effects in both linguistic and musical dimensions, suggesting that participants did also learn the musical structure. The second goal was to evaluate to what extent musical expertise can affect speech segmentation. At this aim, we compared a group of adult musicians to a group of nonmusicians. While behavioral data showed that musicians had marginally better performance than non musicians in both dimensions, electrophysiological data revealed, via early (N1/P2) and late (N400) differences, a better speech segmentation in musicians than in non musicians. Moreover, event-related potentials and time-frequency analyzes during learning revealed a faster and more efficient learning process in musicians. However, the only way to unambiguously claim causality between expertise and the observed effects requires a longitudinal approach. At this aim, we conducted a study with 8 year-old children who followed either music or painting lessons over a period of 2 years. Behavioral and electrophysiological data revealed a larger benefit of musical compared to painting training, bringing evidences for the importance of music in childrens' education
Roques, Martine. „Apprentissage et reconnaissance de structures syntaxiques par une approche connexionniste“. Paris 11, 1993. http://www.theses.fr/1993PA112433.
Der volle Inhalt der QuelleChiapino, Maël. „Apprentissage de structures dans les valeurs extrêmes en grande dimension“. Thesis, Paris, ENST, 2018. http://www.theses.fr/2018ENST0035/document.
Der volle Inhalt der QuelleWe present and study unsupervised learning methods of multivariate extreme phenomena in high-dimension. Considering a random vector on which each marginal is heavy-tailed, the study of its behavior in extreme regions is no longer possible via usual methods that involve finite means and variances. Multivariate extreme value theory provides an adapted framework to this study. In particular it gives theoretical basis to dimension reduction through the angular measure. The thesis is divided in two main part: - Reduce the dimension by finding a simplified dependence structure in extreme regions. This step aim at recover subgroups of features that are likely to exceed large thresholds simultaneously. - Model the angular measure with a mixture distribution that follows a predefined dependence structure. These steps allow to develop new clustering methods for extreme points in high dimension
Chiapino, Maël. „Apprentissage de structures dans les valeurs extrêmes en grande dimension“. Electronic Thesis or Diss., Paris, ENST, 2018. http://www.theses.fr/2018ENST0035.
Der volle Inhalt der QuelleWe present and study unsupervised learning methods of multivariate extreme phenomena in high-dimension. Considering a random vector on which each marginal is heavy-tailed, the study of its behavior in extreme regions is no longer possible via usual methods that involve finite means and variances. Multivariate extreme value theory provides an adapted framework to this study. In particular it gives theoretical basis to dimension reduction through the angular measure. The thesis is divided in two main part: - Reduce the dimension by finding a simplified dependence structure in extreme regions. This step aim at recover subgroups of features that are likely to exceed large thresholds simultaneously. - Model the angular measure with a mixture distribution that follows a predefined dependence structure. These steps allow to develop new clustering methods for extreme points in high dimension
Sablayrolles, Alexandre. „Mémorisation et apprentissage de structures d'indexation avec les réseaux de neurones“. Thesis, Université Grenoble Alpes, 2020. https://thares.univ-grenoble-alpes.fr/2020GRALM044.pdf.
Der volle Inhalt der QuelleMachine learning systems, and in particular deep neural networks, aretrained on large quantities of data. In computer vision for instance, convolutionalneural networks used for image classification, scene recognition,and object detection, are trained on datasets which size ranges from tensof thousands to billions of samples. Deep parametric models have a largecapacity, often in the order of magnitude of the number of datapoints.In this thesis, we are interested in the memorization aspect of neuralnetworks, under two complementary angles: explicit memorization,i.e. memorization of all samples of a set, and implicit memorization,that happens inadvertently while training models. Considering explicitmemorization, we build a neural network to perform approximate setmembership, and show that the capacity of such a neural network scaleslinearly with the number of data points. Given such a linear scaling, weresort to another construction for set membership, in which we build aneural network to produce compact codes, and perform nearest neighborsearch among the compact codes, thereby separating “distribution learning”(the neural network) from storing samples (the compact codes), theformer being independent of the number of samples and the latter scalinglinearly with a small constant. This nearest neighbor system performs amore generic task, and can be plugged in to perform set membership.In the second part of this thesis, we analyze the “unintended” memorizationthat happens during training, and assess if a particular data pointwas used to train a model (membership inference). We perform empiricalmembership inference on large networks, on both individual and groupsof samples. We derive the Bayes-optimal membership inference, andconstruct several approximations that lead to state-of-the-art results inmembership attacks. Finally, we design a new technique, radioactive data,that slightly modifies datasets such that any model trained on them bearsan identifiable mark
Lee, Yun-Ann. „Modélisation de structures syntaxiques complexes pour apprentissage de langue sur le réseau“. Besançon, 2006. http://www.theses.fr/2006BESA1025.
Der volle Inhalt der QuelleThe object of this research is to establish an internet accessible French syntax e-learning system especially for Chinese university students who have studied French and prepare to make further improvement on French syntax. In this dissertation, the Tree Adjoining Grammar (TAG) modularity is applied to evaluation of syntax key differences between French and Chinese and to design of our system. The subordinate clause, one of the most difficult part for Chinese students in French learning plays a major role in this analysis. A general comparative study on Chinese and French syntax is presented first. Two subjects are highlighted : the syntactic differences between two languages as well as problems of Chinese students when they write subordinate clauses. A comparative classification of French and Chinese syntactic structures is presented in two ways: one is examination of French simple phrases with verbal heads. All French-Chinese examples and the process of trees combination are demonstrated in TAG trees. The other is typology of French and Chinese subordination, classified according to the measurement of structure differences between two languages. So the first types are the same, and the last type assembles the sentences with the most significant dissimilar characteristics between French and Chinese. All types are presented in TAG trees. The final step of our research is to create a progressive e-learning system, which is divided into three parts: the first is 11 TAG lessons, the second is French simple phrases with verbal heads in 18 lessons and the last is French subordinate sentences in 25 lessons. The results of survey and test of applicants in the end of this dissertation demonstrate that the proposed approach (TAG) can adapt the e-learning and can further be used to improve the content, the progressions and the planning of our lessons in the future research
Allègre, Olivier. „Adapting the Prerequisite Structure to the Learner in Student Modeling“. Electronic Thesis or Diss., Sorbonne université, 2023. http://www.theses.fr/2023SORUS116.
Der volle Inhalt der QuelleData-driven learner models aim to represent and understand students' knowledge and other meta-cognitive characteristics to support their learning by making predictions about their future performance. Learner modeling can be approached using various complex system models, each providing a different perspective on the student and the learning process. Knowledge-enhanced machine learning techniques, such as Bayesian networks, are particularly well suited for incorporating domain knowledge into the learner model, making them a valuable tool in student modeling.This work explores the modeling and the potential applications of a new framework, called E-PRISM, for Embedding Prerequisite Relationships In Student Modeling, which includes a learner model based on dynamic Bayesian networks. It uses a new architecture for Bayesian networks that rely on the clause of Independence of Causal Influences (ICI), which reduces the number of parameters in the network and allows enhanced interpretability. The study examines the strengths of E-PRISM, including its ability to consider the prerequisite structure between knowledge components, its limited number of parameters, and its enhanced interpretability. The study also introduces a novel approach for approximate inference in large ICI-based Bayesian networks, as well as a performant parameter learning algorithm in ICI-based Bayesian networks. Overall, the study demonstrates the potential of E-PRISM as a promising tool for discovering the prerequisite structure of domain knowledge that may be adapted to the learner with the perspective of improving the outer-loop adaptivity
Poulin-Charronnat, Bénédicte. „Effet d'expertise sur le traitement des structures musicales“. Dijon, 2003. http://www.theses.fr/2003DIJOL012.
Der volle Inhalt der QuelleThe purpose of the present thesis is to evaluate how the perception of musical structures changes according to the expertise of the listeners. A first set of four experiments using the harmonic priming paradigm shows that the listeners, musicians and non-musicians, behave in an identical way in response to the factors manipulated, and notably that they are more sensitive to the cognitive than to the acoustic manipulations. This first result demonstrates that the listeners possess knowledge of the Western tonal system, and independently of their degree of expertise. A fifth study has then specifically evaluated the musical expertise effect. Once again, an absence of difference between musicians and non-musicians suggest that not only music is something acquired by implicit learning procedures but in addition that the perception of music implies cognitive processes that are themselves implicit. In a sixth study, the implicit learning of regularities underlying a new musical system was shown in both musicians and non-musicians. It appears that the implicit characteristic of learning and processing of music are similar to those of language. The two last studies have shown interactions between language and music, suggesting that our brain, by parsimony, could have identical cognitive processes to treat information coming from similar systems
Liquière, Michel. „Apprentissage à partir d'objets structurés : conception et réalisation“. Montpellier 2, 1990. http://www.theses.fr/1990MON20038.
Der volle Inhalt der QuelleSeghouane, Abd-Krim. „Choix de structures de modèles pour traitement robuste“. Paris 11, 2002. http://www.theses.fr/2002PA112244.
Der volle Inhalt der QuelleParametric model identification is an important issue in various research areas like automatic, signal processing, economy and statistics. Defined as the mathematical description of a process from a set of empirical data, model identification have been treated from a theoretical view point by numerous authors and it has application in many practical areas. In this work, we have been interested in the robustness property of the parametrical model used for identification, and in the robustness of the identification procedures. This leads us to consider differently some hypotheses that are generally made in this area. In a first time, parametrical estimation methods have been developed. The error in variables model has been used to construct two parametrical estimation procedures that guaranty the robustness to noise on the experimental inputs. The robustness to noise on the operating inputs has been reviewed in different form and discussed. A means that guaranty the efficiency of the estimation procedures dedicated to this kind of robustness has been proposed. In a second time, we have oriented our interest to the model selection problem. The model selection method that lie on the use of the robustness to noise on the operating inputs property has been reviewed and discussed, and an amelioration has been proposed. The AIC criterion derivation philosophy has been used to construct and to propose another model selection criterion that is more robust for a small sample set. This philosophy is also used to construct a new model selection criterion that is more robust for sample set with missing data
Pitiot, Alain. „Segmentation Automatique des Structures Cérébrales s'appuyant sur des Connaissances Explicites“. Phd thesis, École Nationale Supérieure des Mines de Paris, 2003. http://pastel.archives-ouvertes.fr/pastel-00001346.
Der volle Inhalt der QuelleLeclerc, Sarah Marie-Solveig. „Automatisation de la segmentation sémantique de structures cardiaques en imagerie ultrasonore par apprentissage supervisé“. Thesis, Lyon, 2019. http://www.theses.fr/2019LYSEI121.
Der volle Inhalt der QuelleThe analysis of medical images plays a critical role in cardiology. Ultrasound imaging, as a real-time, low cost and bed side applicable modality, is nowadays the most commonly used image modality to monitor patient status and perform clinical cardiac diagnosis. However, the semantic segmentation (i.e the accurate delineation and identification) of heart structures is a difficult task due to the low quality of ultrasound images, characterized in particular by the lack of clear boundaries. To compensate for missing information, the best performing methods before this thesis relied on the integration of prior information on cardiac shape or motion, which in turns reduced the adaptability of the corresponding methods. Furthermore, such approaches require man- ual identifications of key points to be adapted to a given image, which makes the full process difficult to reproduce. In this thesis, we propose several original fully-automatic algorithms for the semantic segmentation of echocardiographic images based on supervised learning ap- proaches, where the resolution of the problem is automatically set up using data previously analyzed by trained cardiologists. From the design of a dedicated dataset and evaluation platform, we prove in this project the clinical applicability of fully-automatic supervised learning methods, in particular deep learning methods, as well as the possibility to improve the robustness by incorporating in the full process the prior automatic detection of regions of interest
Déjean, Hervé. „Concepts et algorithmes pour la découverte des structures formelles des langues“. Phd thesis, Université de Caen, 1998. http://tel.archives-ouvertes.fr/tel-00169572.
Der volle Inhalt der QuelleVoilà la question à laquelle nous avons essayé de répondre. Cette réponse peut être vue comme une continuation des travaux en analyse distributionnelle développée par Zellig Harris.
L'objectif de ce travail est donc de découvrir les structures formelles d'une langue en étudiant ces régularités formelles contenues dans un corpus
Notre méthode de découverte se base sur une simple conception formelle de la langue: un objet linéaire dans lequel les frontières (de début et de fin) des différentes structures sont indiquées par des éléments caractéristiques. Les structures ainsi identifiées sont le syntagme simple (non récursif), et la proposition, structures à la fois multilingues et formelles. Ces indicateurs de frontières correspondent à des morphèmes (libres ou liés) pour le syntagme, et à des morphèmes ou des syntagmes pour la proposition.
À partir de ces structures théoriques, nous construisons la liste de toutes les catégories qu'un élément (morphème ou mot) peut prendre. Une fois ces structures et catégories recensées, nous construisons des contextes spécifiques à chaque catégorie afin de catégoriser les éléments du texte. Nous obtenons donc un mécanisme permettant d'assigner à un élément plusieurs catégories si cet élément apparaît dans différents contextes. Ces contextes sont construits à l'aide des éléments prototypiques de marqueurs de frontières de structures, identifiables grâce à leur position par rapport à la segmentation physique du texte (en particulier les ponctuations).
Les résultats obtenus permettent la catégorisation des mots du corpus, ainsi qu'une segmentation partielle en syntagmes. La méthode a été appliquée à une dizaine de langues comme le français, l'allemand, le turc, le vietnamien et le swahili.
Morard, Vincent. „Détection de structures fines par traitement d'images et apprentissage statistique : application au contrôle non destructif“. Phd thesis, Ecole Nationale Supérieure des Mines de Paris, 2012. http://pastel.archives-ouvertes.fr/pastel-00932727.
Der volle Inhalt der QuelleNair, Benrekia Noureddine Yassine. „Classification interactive multi-label pour l’aide à l’organisation personnalisée des données“. Nantes, 2015. https://archive.bu.univ-nantes.fr/pollux/show/show?id=bb2e3d25-7f53-4b66-af04-a9fb5e80ea28.
Der volle Inhalt der QuelleThe growing importance given today to personalized contents led to the development of several interactive classification systems for various novel applications. Nevertheless, all these systems use a single-label item classification which greatly constrains the user's expressiveness. The major problem common to all developers of an interactive multi-label system is: which multi-label classifier should we choose? Experimental evaluations of recent interactive learning systems are mainly subjective. The importance of their conclusions is consequently limited. To draw more general conclusions for guiding the selection of a suitable learning algorithm during the development of such a system, we extensively study the impact of the major interactivity constraints (learning from few examples in a limited time) on the classifier predictive and time-computation performances. The experiments demonstrate the potential of an ensemble learning approach Random Forest of Predictive Clustering Trees(RF-PCT). However,the strong constraint imposed by the interactivity on the computation time has led us to propose a new hybrid learning approach FMDI-RF+ which associates RF-PCT with an efficient matrix factorization approach for dimensionality reduction. The experimental results indicate that RF-FMDI+ is as accurate as RF-PCT in the predictions with a significant advantage to FMDI-RF + for the speed of computation
Daucé, Emmanuel. „Adaptation dynamique et apprentissage dans les réseaux de neurones récurrents aléatoires“. Toulouse, ENSAE, 2000. https://tel.archives-ouvertes.fr/tel-01394004.
Der volle Inhalt der QuelleGhrib, Meriem. „Contrôle santé des structures composites : génération de délaminages par choc laser et quantification par apprentissage machine“. Thesis, Paris, ENSAM, 2017. http://www.theses.fr/2017ENAM0070/document.
Der volle Inhalt der QuelleIn this work, we approach delamination quantification in Carbon Fiber Reinforced Polymer (CFRP) laminates as a classification problem whereby each class corresponds to a certain damage extent. A Support Vector Machine (SVM) is used to perform multi-class classification task. Classically, Signal Based Features (SBF) are used to train SVMs when approaching SHM from a machine learning perspective. In this work, starting from the assumption that damage causes a structure to exhibit nonlinear response, we investigate whether the use of Nonlinear Model Based Features (NMBF) increases classification performance. NMBF are computed based on parallel Hammerstein models which are identified with an Exponential Sine Sweep (ESS) signal. Dimensionality reduction of features vector using Principal Component Analysis (PCA) is also conducted in order to find out if it allows robustifying the quantification process suggested in this work. The proposed quantification approach was first tested and validated using simulation results. Thereafter, experimental results on CFRP composite plates equipped with piezoelectric elements and containing various delamination severities are considered for demonstration. Delamination-type damage is introduced into samples in a calibrated and realistic way using LASER Shock Wave Technique (LSWT) and more particularly symmetrical LASER shock configuration. We have experimentally demonstrated that such a configuration of LASER shock is an effective alternative to conventional damage generation techniques such as conventional impacts and Teflon inserts since it allows for a better calibration of damage in type, depth and size
Muller, Jean-Denis. „La perception structurante : apprentissage non monotone de fonctions visuelles par croissance et maturation de structures neuromimétiques“. Toulouse, ENSAE, 1993. http://www.theses.fr/1993ESAE0030.
Der volle Inhalt der QuelleBoyer, Laurent. „Apprentissage probabiliste de similarités d'édition“. Phd thesis, Université Jean Monnet - Saint-Etienne, 2011. http://tel.archives-ouvertes.fr/tel-00718835.
Der volle Inhalt der QuelleCailliès, Stéphanie. „Connaissances initiales, structures textuelles et mémorisation : expérimentations et simulations“. Aix-Marseille 1, 1998. http://www.theses.fr/1998AIX10053.
Der volle Inhalt der QuelleThis research investigates the effect of prior knowledge on comprehension and memorization of expository texts describing the use of softwares. We theorize that readers of different levels of prior knowledge have different knowledge structures, temporal-causal for beginners and teleological for advanced, and that these structures determine the reading times of the textual information to be encoded, implicit and explicit, and the retrieval of this information. More precisely, we assume that for advanced, the goal and the outcome are directly and strongly associated in memory while the actions, necessary to attain the goal, are subordinated, and that the type of relationships beginners established among the goal, the outcome, and the actions varies according to the order in which they appear in the text. Six experimentations were realized to test this hypothesis. The first test was to show a comprehension facilitation ascribed to the homology of structures between prior knowledge and texts. The second test was to use a primed recognition task that allowed us to study the relationship readers with different levels of prior knowledge established among the goal of a sequence, the actions necessary to attain the goal, and the obtained outcome. We assume that the relationship advanced established among the goal, the actions and the outcome depends on their prior knowledge structures whereas that of beginners varies according to the order in which textual information appears in the text. Four experiments, distinguished by different presentation orders of sentences expressed a goal, an action and an outcome, were realized. The results, that support our main assumption, were simulated and reproduced with the construction-integration model proposed by kintsch (1988, 1998)
Héas, Patrick. „Apprentissage bayésien de structures spatio-temporelles : application à la fouille visuelle de séries temporelles d'images de satellites“. Toulouse, ENSAE, 2005. http://www.theses.fr/2005ESAE0004.
Der volle Inhalt der QuelleLatchoumanin, Michel. „Langue maternelle et apprentissage cognitif : apport d'une expérience d'induction de structures cognitives auprès des jeunes enfants réunionnais“. Aix-Marseille 1, 1991. http://www.theses.fr/1991AIX10062.
Der volle Inhalt der QuelleWithin a neo-structuralist approach of cognitive development, which attemps to organise a piagetian genetic psychology with an information processing theory, we tried to analyse the processes in young children from reunion island reared cognitive inferred from the induction referential reasoning within a diglossic context. The existence and the nature of the processes involved are inferred from verbal behaviours preceding and following actions in cognitive training situation. Gains are identified according to developmental steps and modes of representations which are genuine to the pre-operative period. The analysis of the results gives strong evidence of significant cognitive improvements when the Creole mother tongue is used with subjects not well acquainted with the official teaching language (French). The experiment shows that the children have acquired significant competences in cognitive situations far remote from those usually obtained in traditional school training in French, this at an earlier age and with better results than those obtained with children taught in the official language, or not taught at all. More over, these competences seem also to appear in context situations different from the training situations these results bear a significant statistical difference the two groups experimental (training in French and in Creole) and the control group (no training)
Blin, Laurent. „Apprentissage de structures d'arbres à partir d'exemples ; application à la prosodie pour la synthèse de la parole“. Rennes 1, 2002. http://www.theses.fr/2002REN10117.
Der volle Inhalt der QuelleHoch, Lisianne. „Perception et apprentissage des structures musicales et langagières : études des ressources cognitives partagées et des effets attentionnels“. Thesis, Lyon 2, 2010. http://www.theses.fr/2010LYO20049/document.
Der volle Inhalt der QuelleMusic and language are structurally organized materials that are based on combinatorial principles. Listeners have acquired knowledge about these structural regularities via mere exposure. This knowledge allows them to develop expectations about future events in music and language perception. My PhD investigated two aspects of domain-specificity versus generality of cognitive functions in music and language processing: perception and statistical learning.In the first part (perception), musical structure processing has been shown to influence spoken and visual language processing (Études 1 & 4), partly due to dynamic attending mechanisms (Jones, 1976). More specifically, musical structure processing has been shown to interact with linguistic-syntactic processing, but not with linguistic-semantic processing (Étude 3), thus supporting the hypothesis of shared syntactic resources for music and language processing (Patel, 2003). Together with previous studies that have investigated simultaneous musical and linguistic (syntactic and semantic) structure processing, we proposed that these shared resources might extend to the processing of other structurally organized information that require structural and temporal integration resources. This hypothesis was tested and supported by interactive influences between simultaneous musical and arithmetic structure processing (Étude 4). In the second part (learning), statistical learning was directly compared for verbal and nonverbal materials. In particular, we aimed to investigate the influence of dynamic attention driven by non-acoustic (Études 5 & 6) and acoustic (Étude 7) cues on statistical learning. Non-acoustic temporal cues have been shown to influence statistical learning of verbal and nonverbal artificial languages. In agreement with the dynamic attending theory (Jones, 1976), we proposed that non-acoustic temporal cues guide attention over time and influence statistical learning.Based on the influence of dynamic attending mechanisms on perception and learning and on evidence of shared structural and temporal integration resources for the processing of musical structures and other structured information, this PhD opens new questions about the potential influence of tonal and temporal auditory structure processing on general cognitive sequencing abilities, notably required in structured sequence perception and learning.Jones, M. R. (1976). Time, our lost dimension: Toward a new theory of perception, attention, and memory. Psychological Review, 83(5), 323-355. doi:10.1037/0033-295X.83.5.323Patel, A. D. (2003). Language, music, syntax and the brain. Nature Neuroscience, 6(7), 674-681. doi:10.1038/nn1082
Bouabdallah, Khaled. „Structures d'emploi, filieres industrielles et competitivite. Essai sur le travail et la performance economique“. Lyon 2, 1993. http://www.theses.fr/1993LYO22005.
Der volle Inhalt der QuelleFrom an analysis of the links between labour and economic performance, we have shown the need for renewal of problematic s. As a matter of fact, we have to identify the elements founding the conditions of the contemporary competitiveness and the renewed mechanisms of growth, these latter depending more and more broadly on endogeneous factors. The analytical framework provided is rooted in a conception of competitiveness enabling to take into account externalities, technologic al interdependences and learning effects, and more largely the structural factors ensuring a long-lastin competitiveness. The meso-economic seems to provide a relevant way for expresing such phenomena. Then the interest of the level of the filiere is learning processes and externalities. This level enables also to integrate concerns linked to the analysis o f labour. On this basis, an applied analysis is carried out about industrial filieres, wich the aim is both to investiga te, on the medium run, the dynamic of the filiere employment structures, and to confront this latter with the evolution of their economic performance. Phenomena. Then the interest of the level of the filiere is emphasized, inasmuch it enables to express interdependence effects, learning processes and externalities. This level enables also to integrate concerns linked to the analysis of labour. On this basis, an applied analysis is carried out about industrial filieres, wich the aim is both to investigate, on the medium run, the dynamic of the filiere employment structures, and to confront this latter with the evolution of their economic performance
Jeong, Seong-Gyun. „Modélisation de structures curvilignes et ses applications en vision par ordinateur“. Thesis, Nice, 2015. http://www.theses.fr/2015NICE4086/document.
Der volle Inhalt der QuelleIn this dissertation, we propose curvilinear structure reconstruction models based on stochastic modeling and ranking learning system. We assume that the entire line network can be decomposed into a set of line segments with variable lengths and orientations. This assumption enables us to reconstruct arbitrary shapes of curvilinear structure for different types of datasets. We compute curvilinear feature descriptors based on the image gradient profiles and the morphological profiles. For the stochastic model, we propose prior constraints that define the spatial interaction of line segments. To obtain an optimal configuration corresponding to the latent curvilinear structure, we combine multiple line hypotheses which are computed by MCMC sampling with different parameter sets. Moreover, we learn a ranking function which predicts the correspondence of the given line segment and the latent curvilinear structures. A novel graph-based method is proposed to infer the underlying curvilinear structure using the output rankings of the line segments. We apply our models to analyze curvilinear structure on static images. Experimental results on wide types of datasets demonstrate that the proposed curvilinear structure modeling outperforms the state-of-the-art techniques
Poulenard, Adrien. „Structures for deep learning and topology optimization of functions on 3D shapes“. Thesis, Institut polytechnique de Paris, 2020. http://www.theses.fr/2020IPPAX007.
Der volle Inhalt der QuelleThe field of geometry processing is following a similar path as image analysis with the explosion of publications dedicated to deep learning in recent years. An important research effort is being made to reproduce the successes of deep learning 2D computer vision in the context of 3D shape analysis. Unlike images shapes comes in various representations like meshes or point clouds which often lack canonical structure. This makes traditional deep learning algorithms like Convolutional Neural Networks (CNN) non straightforward to apply to 3D data. In this thesis we propose three main contributions:First, we introduce a method to compare functions on different domains without correspondences and to deform them to make the topology of their set of levels more alike. We apply our method to the classical problem of shape matching in the context of functional maps to produce smoother and more accurate correspondences. Furthermore, our method is based on the continuous optimization of a differentiable energy with respect to the compared functions and is applicable to deep learning. We make two direct contributions to deep learning on 3D data. We introduce a new convolution operator over triangles meshes based on local polar coordinates and apply it to deep learning on meshes. Unlike previous works our operator takes all choices of polar coordinates into account without loss of directional information. Lastly we introduce a new rotation invariant convolution layer over point clouds and show that CNNs based on this layer can outperform state of the art methods in standard tasks on un-alligned datasets even with data augmentation
Zotti, Clément. „Réseaux de neurones à convolutions pour la segmentation multi structures d'images par résonance magnétique cardiaque“. Mémoire, Université de Sherbrooke, 2018. http://hdl.handle.net/11143/11817.
Der volle Inhalt der QuelleBesombes, Jérôme. „Un modèle algorithmique de la généralisation de structures dans le processus d'acquisition du langage“. Nancy 1, 2003. http://www.theses.fr/2003NAN10156.
Der volle Inhalt der QuelleThe subject of our study is the learning of regular tree languages for an algorithmic modeling of language acquisition. For this, we suppose that data are structured; these data are heard correct sentences and the learning is effective since a representation of the language to which these sentences belong is built. From this representation the learner is able to generate new sentences compatible with the language and not presented as examples. Considering that heard sentences are translated into trees, it appears that the generalization of these tree structures is a component of the learning. We developed several models for this generalization in the form of algorithms taking into account various types of structures as input and various levels of contribution of information. These new models offer the advantage of unifying major results in the theory of the grammatical inference, and of extending these results, in particular by the consideration of new structures not studied previously in the learnability point of view
Oyallon, Edouard. „Analyzing and introducing structures in deep convolutional neural networks“. Thesis, Paris Sciences et Lettres (ComUE), 2017. http://www.theses.fr/2017PSLEE060.
Der volle Inhalt der QuelleThis thesis studies empirical properties of deep convolutional neural networks, and in particular the Scattering Transform. Indeed, the theoretical analysis of the latter is hard and until now remains a challenge: successive layers of neurons have the ability to produce complex computations, whose nature is still unknown, thanks to learning algorithms whose convergence guarantees are not well understood. However, those neural networks are outstanding tools to tackle a wide variety of difficult tasks, like image classification or more formally statistical prediction. The Scattering Transform is a non-linear mathematical operator whose properties are inspired by convolutional networks. In this work, we apply it to natural images, and obtain competitive accuracies with unsupervised architectures. Cascading a supervised neural networks after the Scattering permits to compete on ImageNet2012, which is the largest dataset of labeled images available. An efficient GPU implementation is provided. Then, this thesis focuses on the properties of layers of neurons at various depths. We show that a progressive dimensionality reduction occurs and we study the numerical properties of the supervised classification when we vary the hyper parameters of the network. Finally, we introduce a new class of convolutional networks, whose linear operators are structured by the symmetry groups of the classification task
Maag, Maria Coralia Laura. „Apprentissage automatique de fonctions d'anonymisation pour les graphes et les graphes dynamiques“. Electronic Thesis or Diss., Paris 6, 2015. http://www.theses.fr/2015PA066050.
Der volle Inhalt der QuelleData privacy is a major problem that has to be considered before releasing datasets to the public or even to a partner company that would compute statistics or make a deep analysis of these data. Privacy is insured by performing data anonymization as required by legislation. In this context, many different anonymization techniques have been proposed in the literature. These techniques are difficult to use in a general context where attacks can be of different types, and where measures are not known to the anonymizer. Generic methods able to adapt to different situations become desirable. We are addressing the problem of privacy related to graph data which needs, for different reasons, to be publicly made available. This corresponds to the anonymized graph data publishing problem. We are placing from the perspective of an anonymizer not having access to the methods used to analyze the data. A generic methodology is proposed based on machine learning to obtain directly an anonymization function from a set of training data so as to optimize a tradeoff between privacy risk and utility loss. The method thus allows one to get a good anonymization procedure for any kind of attacks, and any characteristic in a given set. The methodology is instantiated for simple graphs and complex timestamped graphs. A tool has been developed implementing the method and has been experimented with success on real anonymized datasets coming from Twitter, Enron or Amazon. Results are compared with baseline and it is showed that the proposed method is generic and can automatically adapt itself to different anonymization contexts
Pesquerel, Fabien. „Information per unit of interaction in stochastic sequential decision making“. Electronic Thesis or Diss., Université de Lille (2022-....), 2023. https://pepite-depot.univ-lille.fr/LIBRE/EDMADIS/2023/2023ULILB048.pdf.
Der volle Inhalt der QuelleIn this thesis, we wonder about the rate at which one can solve an unknown stochastic problem.To this purpose we introduce two research fields known as Bandit and Reinforcement Learning.In these two settings, a learner must sequentially makes decision that will affect a reward signal that the learner receive.The learner does not know the environment with which it is interaction, yet wish to maximize its average reward in the long run.More specifically, we are interested in studying some form of stochastic decision problem under the average-reward criterion in which a learning algorithm interacts sequentially with a dynamical system, without any reset, in a single and infinite sequence of observations, actions, and rewards while trying to maximize its total accumulated rewards over time.We first introduce Bandit, in which the set of decision is constant and introduce what is meant by solving the problem.Amongst those learners, some are better than all the others, and called optimal.We first focus on how to make the most out of each interaction with the system by revisiting an optimal algorithm, and reduce its numerical complexity.Therefore, the information extracted from each sample, per-time-step, is larger since the optimality remains.Then we study an interesting structured problem in which one can exploit the structure without estimating it.Afterward we introduce Reinforcement Learning, in which the decision a learner can make depend on a notion of state.Each time a learner makes a decision, it receives a reward and the state change according to transition law on the set of states.In some setting, known as ergodic, an optimal rate of solving is known and we introduce a knew algorithm that we can prove to be optimal and show to be numerically efficient.In a final chapter, we make a step in the direction of removing the ergodic assumption by considering the a priori simpler problem where the transitions are known.Yet, correctly understanding the rate at which information can be acquired about an optimal solution is already not easy
Fauconnier, Jean-Philippe. „Acquisition de liens sémantiques à partir d'éléments de mise en forme des textes : exploitation des structures énumératives“. Thesis, Toulouse 3, 2016. http://www.theses.fr/2016TOU30023.
Der volle Inhalt der QuelleThe past decade witnessed significant advances in the field of relation extraction from text, facilitating the building of lexical or semantic resources. However, the methods proposed so far (supervised learning, kernel methods, distant supervision, etc.) don't fully exploit the texts : they are usually applied at the sentential level and they don't take into account the layout and the formatting of texts. In such a context, this thesis aims at expanding those methods and makes them layout-aware for extracting relations expressed beyond sentence boundaries. For this purpose, we rely on the semantics conveyed by typographical (bullets, emphasis, etc.) and dispositional (visual indentations, carriage returns, etc.) features. Those features often substitute purely discursive formulations. In particular, the study reported here is dealing with the relations carried by the vertical enumerative structures. Although they display discontinuities between their various components, the enumerative structures can be dealt as a whole at the semantic level. They form textual structures prone to hierarchical relations. This study was divided into two parts. (i) The first part describes a model representing the hierarchical structure of documents. This model is falling within the theoretical framework representing the textual architecture : an abstraction of the layout and the formatting, as well as a strong connection with the rhetorical structure are achieved. However, our model focuses primarily on the efficiency of the analysis process rather than on the expressiveness of the representation. A bottom-up method intended for building automatically this model is presented and evaluated on a corpus of PDF documents. (ii) The second part aims at integrating this model into the process of relation extraction. In particular, we focused on vertical enumerative structures. A multidimensional typology intended for characterizing those structures was established and used into an annotation task. Thanks to corpus-based observations, we proposed a two-step method, by supervised learning, for qualifying the nature of the relation and identifying its arguments. The evaluation of our method showed that exploiting the formatting and the layout of documents, in combination with standard lexico-syntactic features, improves those two tasks
Bouthinon, Dominique. „Apprentissage à partir d'exemples ambigus : étude théorique et application à la découverte de structures communes à un ensemble de séquences d'ARN“. Paris 13, 1996. http://www.theses.fr/1996PA132033.
Der volle Inhalt der QuelleRan, Peipei. „Imaging and diagnostic of sub-wavelength micro-structures, from closed-form algorithms to deep learning“. Electronic Thesis or Diss., université Paris-Saclay, 2020. http://www.theses.fr/2020UPASG061.
Der volle Inhalt der QuelleElectromagnetic probing of a gridlike, finite set of infinitely long circular cylindrical dielectric rods affected by missing ones is investigated from time-harmonic single and multiple frequency data. Sub-wavelength distances between adjacent rods and sub-wavelength rod diameters are assumed throughout the frequency band of operation and this leads to a severe challenge due to need of super-resolution within the present micro-structure, well beyond the Rayleigh criterion. A wealth of solution methods is investigated and comprehensive numerical simulations illustrate pros and cons, completed by processing laboratory-controlled experimental data acquired on a micro-structure prototype in a microwave anechoic chamber. These methods, which differ per a priori information accounted for and consequent versatility, include time-reversal, binary-specialized contrast-source and sparsity-constrained inversions, and convolutional neural networks possibly combined with recurrent ones
You, Weizhen. „Reliability assessment of TMD-based control structures : A Statistical Learning Perspective“. Thesis, Lyon, 2020. http://www.theses.fr/2020LYSEC001.
Der volle Inhalt der QuelleThe study of structural reliability mainly concerns the evaluation and prediction of the risk of limit state violation for an engineering structure at any stage of its life. Reliability evaluation helps improve structure design and product quality, which is of great significance for companies and consumers. It is also the basis of reliability modeling and prediction. Vibration control is a technique to reduce the energy of a vibrating structure when it is excited by external forces. This technique is widely used in various systems, such as buildings, bridges, machine tools and vehicles. Reliability prediction helps companies make production planning and implement preventive maintenance. To do the predictions, a reliability model is firstly determined. Due to complex interior and exterior factors, the structure properties always deviate their design values. The structural uncertainties play an important role in reliability modeling. Traditional reliability models are commonly based on a priori information and professional knowledge, which has been unrealistic for today’s systems that are more complex and nonlinear due to advanced design methodologies. In this situation, growing attention has been paid to non-parametric statistical learning approaches. Seen as a classification/ regression procedure, the prediction task can be realized by machine learning models, such as Tree methods, Support vector machines, Artificial Neural Networks, etc. These models are attracting more and more attention in recent published researches. In this research, we have investigated several machine learning models such as Random Forests, Adaptive Boosting, Support vector machines, Artificial Neural Networks, etc. Besides we developed a new system reliability assessment method for complex structural systems. These methods extend statistical learning methods on structural reliability analysis and prediction
Akindele, Oluwatoyin Tunde. „Vers un système de construction automatique de modèles génériques de structures de documents“. Nancy 1, 1995. http://www.theses.fr/1995NAN10002.
Der volle Inhalt der Quelle