Dissertationen zum Thema „Apprentissage automatique non supervisée“
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 automatique non supervisée" 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.
Delsert, Stéphane. „Classification interactive non supervisée de données multidimensionnelles par réseaux de neurones à apprentissage cométitif“. Lille 1, 1996. https://pepite-depot.univ-lille.fr/LIBRE/Th_Num/1996/50376-1996-214.pdf.
Der volle Inhalt der QuelleGuérif, Sébastien. „Réduction de dimension en apprentissage numérique non supervisé“. Paris 13, 2006. http://www.theses.fr/2006PA132032.
Der volle Inhalt der QuellePeyrache, Jean-Philippe. „Nouvelles approches itératives avec garanties théoriques pour l'adaptation de domaine non supervisée“. Thesis, Saint-Etienne, 2014. http://www.theses.fr/2014STET4023/document.
Der volle Inhalt der QuelleDuring the past few years, an increasing interest for Machine Learning has been encountered, in various domains like image recognition or medical data analysis. However, a limitation of the classical PAC framework has recently been highlighted. It led to the emergence of a new research axis: Domain Adaptation (DA), in which learning data are considered as coming from a distribution (the source one) different from the one (the target one) from which are generated test data. The first theoretical works concluded that a good performance on the target domain can be obtained by minimizing in the same time the source error and a divergence term between the two distributions. Three main categories of approaches are derived from this idea : by reweighting, by reprojection and by self-labeling. In this thesis work, we propose two contributions. The first one is a reprojection approach based on boosting theory and designed for numerical data. It offers interesting theoretical guarantees and also seems able to obtain good generalization performances. Our second contribution consists first in a framework filling the gap of the lack of theoretical results for self-labeling methods by introducing necessary conditions ensuring the good behavior of this kind of algorithm. On the other hand, we propose in this framework a new approach, using the theory of (epsilon, gamma, tau)- good similarity functions to go around the limitations due to the use of kernel theory in the specific context of structured data
Cleuziou, Guillaume. „Une méthode de classification non-supervisée pour l'apprentissage de règles et la recherche d'information“. Phd thesis, Université d'Orléans, 2004. http://tel.archives-ouvertes.fr/tel-00084828.
Der volle Inhalt der QuelleNous proposons, dans cette étude, l'algorithme de clustering PoBOC permettant de structurer un ensemble d'objets en classes non-disjointes. Nous utilisons cette méthode de clustering comme outil de traitement dans deux applications très différentes.
- En apprentissage supervisé, l'organisation préalable des instances apporte une connaissance utile pour la tâche d'induction de règles propositionnelles et logiques.
- En Recherche d'Information, les ambiguïtés et subtilités de la langue naturelle induisent naturellement des recouvrements entre thématiques.
Dans ces deux domaines de recherche, l'intérêt d'organiser les objets en classes non-disjointes est confirmé par les études expérimentales adaptées.
Fischer, Aurélie. „Apprentissage statistique non supervisé : grande dimension et courbes principales“. Paris 6, 2011. http://www.theses.fr/2011PA066142.
Der volle Inhalt der QuelleRibeiro, Swen. „Induction non-supervisée de schémas d’évènements à partir de textes journalistiques“. Thesis, université Paris-Saclay, 2020. http://www.theses.fr/2020UPASS059.
Der volle Inhalt der QuelleEvents are central in many Natural Language Processing tasks, despite the lack of a unified definition for the concept. The field of event processing took off with the MUC evaluation campaigns that provided participants with reference structures called templates. These templates were composed of a title (the name of the event) and several slots, i.e specific and atomic pieces of data about the event. Creating these templates is an expert task and therefore costly, painstaking and hard to extend to new domains.Meanwhile, the amount of data produced by individuals and organizations has grown exponentially, opening unprecedented perspectives of applications. In the journalistic domain, it fueled the development of a new paradigm called data-journalism.In this work, we aim at inducing synthetic representations of events from large textual journalistic corpora. These representations would be comparable to MUC templates and used by data-journalists to explore large textual news datasets. To this end, we propose a bottom-up approach composed of three main steps. The first step clusters several textual mentions of a same particular event (i.e tied to a time and place) to identify distinct instances. The second step groups these instances together based on more abstract features to infer event types. Finally, the third and last step extracts the most salient elements of each type to produce the synthetic, template-like structure we are looking for
Sublemontier, Jacques-Henri. „Classification non supervisée : de la multiplicité des données à la multiplicité des analyses“. Phd thesis, Université d'Orléans, 2012. http://tel.archives-ouvertes.fr/tel-00801555.
Der volle Inhalt der QuelleBach, Tran. „Algorithmes avancés de DCA pour certaines classes de problèmes en apprentissage automatique du Big Data“. Electronic Thesis or Diss., Université de Lorraine, 2019. http://www.theses.fr/2019LORR0255.
Der volle Inhalt der QuelleBig Data has become gradually essential and ubiquitous in all aspects nowadays. Therefore, there is an urge to develop innovative and efficient techniques to deal with the rapid growth in the volume of data. This dissertation considers the following problems in Big Data: group variable selection in multi-class logistic regression, dimension reduction by t-SNE (t-distributed Stochastic Neighbor Embedding), and deep clustering. We develop advanced DCAs (Difference of Convex functions Algorithms) for these problems, which are based on DC Programming and DCA – the powerful tools for non-smooth non-convex optimization problems. Firstly, we consider the problem of group variable selection in multi-class logistic regression. We tackle this problem by using recently advanced DCAs -- Stochastic DCA and DCA-Like. Specifically, Stochastic DCA specializes in the large sum of DC functions minimization problem, which only requires a subset of DC functions at each iteration. DCA-Like relaxes the convexity condition of the second DC component while guaranteeing the convergence. Accelerated DCA-Like incorporates the Nesterov's acceleration technique into DCA-Like to improve its performance. The numerical experiments in benchmark high-dimensional datasets show the effectiveness of proposed algorithms in terms of running time and solution quality. The second part studies the t-SNE problem, an effective non-linear dimensional reduction technique. Motivated by the novelty of DCA-Like and Accelerated DCA-Like, we develop two algorithms for the t-SNE problem. The superiority of proposed algorithms in comparison with existing methods is illustrated through numerical experiments for visualization application. Finally, the third part considers the problem of deep clustering. In the first application, we propose two algorithms based on DCA to combine t-SNE with MSSC (Minimum Sum-of-Squares Clustering) by following two approaches: “tandem analysis” and joint-clustering. The second application considers clustering with auto-encoder (a well-known type of neural network). We propose an extension to a class of joint-clustering algorithms to overcome the scaling problem and applied for a specific case of joint-clustering with MSSC. Numerical experiments on several real-world datasets show the effectiveness of our methods in rapidity and clustering quality, compared to the state-of-the-art methods
Martel-Brisson, Nicolas. „Approche non supervisée de segmentation de bas niveau dans un cadre de surveillance vidéo d'environnements non contrôlés“. Thesis, Université Laval, 2012. http://www.theses.ulaval.ca/2012/29093/29093.pdf.
Der volle Inhalt der QuelleSîrbu, Adela-Maria. „Dynamic machine learning for supervised and unsupervised classification“. Thesis, Rouen, INSA, 2016. http://www.theses.fr/2016ISAM0002/document.
Der volle Inhalt der QuelleThe research direction we are focusing on in the thesis is applying dynamic machine learning models to salve supervised and unsupervised classification problems. We are living in a dynamic environment, where data is continuously changing and the need to obtain a fast and accurate solution to our problems has become a real necessity. The particular problems that we have decided te approach in the thesis are pedestrian recognition (a supervised classification problem) and clustering of gene expression data (an unsupervised classification. problem). The approached problems are representative for the two main types of classification and are very challenging, having a great importance in real life.The first research direction that we approach in the field of dynamic unsupervised classification is the problem of dynamic clustering of gene expression data. Gene expression represents the process by which the information from a gene is converted into functional gene products: proteins or RNA having different roles in the life of a cell. Modern microarray technology is nowadays used to experimentally detect the levels of expressions of thousand of genes, across different conditions and over time. Once the gene expression data has been gathered, the next step is to analyze it and extract useful biological information. One of the most popular algorithms dealing with the analysis of gene expression data is clustering, which involves partitioning a certain data set in groups, where the components of each group are similar to each other. In the case of gene expression data sets, each gene is represented by its expression values (features), at distinct points in time, under the monitored conditions. The process of gene clustering is at the foundation of genomic studies that aim to analyze the functions of genes because it is assumed that genes that are similar in their expression levels are also relatively similar in terms of biological function.The problem that we address within the dynamic unsupervised classification research direction is the dynamic clustering of gene expression data. In our case, the term dynamic indicates that the data set is not static, but it is subject to change. Still, as opposed to the incremental approaches from the literature, where the data set is enriched with new genes (instances) during the clustering process, our approaches tackle the cases when new features (expression levels for new points in time) are added to the genes already existing in the data set. To our best knowledge, there are no approaches in the literature that deal with the problem of dynamic clustering of gene expression data, defined as above. In this context we introduced three dynamic clustering algorithms which are able to handle new collected gene expression levels, by starting from a previous obtained partition, without the need to re-run the algorithm from scratch. Experimental evaluation shows that our method is faster and more accurate than applying the clustering algorithm from scratch on the feature extended data set
Arcadias, Marie. „Apprentissage non supervisé de dépendances à partir de textes“. Thesis, Orléans, 2015. http://www.theses.fr/2015ORLE2080/document.
Der volle Inhalt der QuelleDependency grammars allow the construction of a hierarchical organization of the words of sentences. The one-by-one building of dependency trees can be very long and it requries expert knowledge. In this regard, we are interested in unsupervised dependency learning. Currently, DMV give the state-of-art results in unsupervised dependency parsing. However, DMV has been known to be highly sensitive to initial parameters. The training of DMV model is also heavy and long. We present in this thesis a new model to solve this problem in a simpler, faster and more adaptable way. We learn a family of PCFG using less than 6 nonterminal symbols and less than 15 combination rules from the part-of-speech tags. The tuning of these PCFG is ligth, and so easily adaptable to the 12 languages we tested. Our proposed method for unsupervised dependency parsing can show the near state-of-the-art results, being twice faster. Moreover, we describe our interests in dependency trees to other applications such as relation extraction. Therefore, we show how such information from dependency structures can be integrated into condition random fields and how to improve a relation extraction task
Pessiot, Jean-François. „Apprentissage automatique pour l'extraction de caractéristiques : application au partitionnement de documents, au résumé automatique et au filtrage collaboratif“. Paris 6, 2008. http://www.theses.fr/2008PA066218.
Der volle Inhalt der QuelleLöser, Kevin. „Apprentissage non-supervisé de la morphologie des langues à l’aide de modèles bayésiens non-paramétriques“. Thesis, Université Paris-Saclay (ComUE), 2019. http://www.theses.fr/2019SACLS203/document.
Der volle Inhalt der QuelleA crucial issue in statistical natural language processing is the issue of sparsity, namely the fact that in a given learning corpus, most linguistic events have low occurrence frequencies, and that an infinite number of structures allowed by a language will not be observed in the corpus. Neural models have already contributed to solving this issue by inferring continuous word representations. These continuous representations allow to structure the lexicon by inducing semantic or syntactic similarity between words. However, current neural models only partially solve the sparsity issue, due to the fact that they require a vectorial representation for every word in the lexicon, but are unable to infer sensible representations for unseen words. This issue is especially present in morphologically rich languages, where word formation processes yield a proliferation of possible word forms, and little overlap between the lexicon observed during model training, and the lexicon encountered during its use. Today, several languages are used on the Web besides English, and engineering translation systems that can handle morphologies that are very different from western European languages has become a major stake. The goal of this thesis is to develop new statistical models that are able to infer in an unsupervised fashion the word formation processes underlying an observed lexicon, in order to produce morphological analyses of new unseen word forms
Sicard, Rudy. „Modélisation d'interdépendances et principe de la moyenne Bayésienne des modèles dans le cadre supervisé et non supervisé“. Paris 6, 2008. http://www.theses.fr/2008PA066696.
Der volle Inhalt der QuelleYin, Hao. „Étude des réseaux de neurones en mode non supervisé : application à la reconnaissance des formes“. Compiègne, 1992. http://www.theses.fr/1992COMPD524.
Der volle Inhalt der QuelleYahiaoui, Meriem. „Modèles statistiques avancés pour la segmentation non supervisée des images dégradées de l'iris“. Electronic Thesis or Diss., Université Paris-Saclay (ComUE), 2017. http://www.theses.fr/2017SACLL006.
Der volle Inhalt der QuelleIris is considered as one of the most robust and efficient modalities in biometrics because of its low error rates. These performances were observed in controlled situations, which impose constraints during the acquisition in order to have good quality images. The renouncement of these constraints, at least partially, implies degradations in the quality of the acquired images and it is therefore a degradation of these systems’ performances. One of the main proposed solutions in the literature to take into account these limits is to propose a robust approach for iris segmentation. The main objective of this thesis is to propose original methods for the segmentation of degraded images of the iris. Markov chains have been well solicited to solve image segmentation problems. In this context, a feasibility study of unsupervised segmentation into regions of degraded iris images by Markov chains was performed. Different image transformations and different segmentation methods for parameters initialization have been studied and compared. Optimal modeling has been inserted in iris recognition system (with grayscale images) to produce a comparison with the existing methods. Finally, an extension of the modeling based on the hidden Markov chains has been developed in order to realize an unsupervised segmentation of the iris images acquired in visible light
Martin, Louis. „Simplification automatique de phrases à l'aide de méthodes contrôlables et non supervisées“. Electronic Thesis or Diss., Sorbonne université, 2021. http://www.theses.fr/2021SORUS265.
Der volle Inhalt der QuelleIn this thesis we study the task of automatic sentence simplification. We first study the different methods used to evaluate simplification models, highlight several shortcomings of current approaches, and propose new contributions. We then propose to train sentence simplification models that can be adapted to the target user, allowing for greater simplification flexibility. Finally, we extend the scope of sentence simplification to several languages, by proposing methods that do not require annotated training data, but that nevertheless achieve very strong performance
Doquet, Guillaume. „Agnostic Feature Selection“. Electronic Thesis or Diss., Université Paris-Saclay (ComUE), 2019. http://www.theses.fr/2019SACLS486.
Der volle Inhalt der QuelleWith the advent of Big Data, databases whose size far exceed the human scale are becoming increasingly common. The resulting overabundance of monitored variables (friends on a social network, movies watched, nucleotides coding the DNA, monetary transactions...) has motivated the development of Dimensionality Reduction (DR) techniques. A DR algorithm such as Principal Component Analysis (PCA) or an AutoEncoder typically combines the original variables into new features fewer in number, such that most of the information in the dataset is conveyed by the extracted feature set.A particular subcategory of DR is formed by Feature Selection (FS) methods, which directly retain the most important initial variables. How to select the best candidates is a hot topic at the crossroad of statistics and Machine Learning. Feature importance is usually inferred in a supervised context, where variables are ranked according to their usefulness for predicting a specific target feature.The present thesis focuses on the unsupervised context in FS, i.e. the challenging situation where no prediction goal is available to help assess feature relevance. Instead, unsupervised FS algorithms usually build an artificial classification goal and rank features based on their helpfulness for predicting this new target, thus falling back on the supervised context. Additionally, the efficiency of unsupervised FS approaches is typically also assessed in a supervised setting.In this work, we propose an alternate model combining unsupervised FS with data compression. Our Agnostic Feature Selection (AgnoS) algorithm does not rely on creating an artificial target and aims to retain a feature subset sufficient to recover the whole original dataset, rather than a specific variable. As a result, AgnoS does not suffer from the selection bias inherent to clustering-based techniques.The second contribution of this work( Agnostic Feature Selection, G. Doquet & M. Sebag, ECML PKDD 2019) is to establish both the brittleness of the standard supervised evaluation of unsupervised FS, and the stability of the new proposed AgnoS
Allesiardo, Robin. „Bandits Manchots sur Flux de Données Non Stationnaires“. Thesis, Université Paris-Saclay (ComUE), 2016. http://www.theses.fr/2016SACLS334/document.
Der volle Inhalt der QuelleThe multi-armed bandit is a framework allowing the study of the trade-off between exploration and exploitation under partial feedback. At each turn t Є [1,T] of the game, a player has to choose an arm kt in a set of K and receives a reward ykt drawn from a reward distribution D(µkt) of mean µkt and support [0,1]. This is a challeging problem as the player only knows the reward associated with the played arm and does not know what would be the reward if she had played another arm. Before each play, she is confronted to the dilemma between exploration and exploitation; exploring allows to increase the confidence of the reward estimators and exploiting allows to increase the cumulative reward by playing the empirical best arm (under the assumption that the empirical best arm is indeed the actual best arm).In the first part of the thesis, we will tackle the multi-armed bandit problem when reward distributions are non-stationary. Firstly, we will study the case where, even if reward distributions change during the game, the best arm stays the same. Secondly, we will study the case where the best arm changes during the game. The second part of the thesis tacles the contextual bandit problem where means of reward distributions are now dependent of the environment's current state. We will study the use of neural networks and random forests in the case of contextual bandits. We will then propose meta-bandit based approach for selecting online the most performant expert during its learning
Yahiaoui, Meriem. „Modèles statistiques avancés pour la segmentation non supervisée des images dégradées de l'iris“. Thesis, Université Paris-Saclay (ComUE), 2017. http://www.theses.fr/2017SACLL006/document.
Der volle Inhalt der QuelleIris is considered as one of the most robust and efficient modalities in biometrics because of its low error rates. These performances were observed in controlled situations, which impose constraints during the acquisition in order to have good quality images. The renouncement of these constraints, at least partially, implies degradations in the quality of the acquired images and it is therefore a degradation of these systems’ performances. One of the main proposed solutions in the literature to take into account these limits is to propose a robust approach for iris segmentation. The main objective of this thesis is to propose original methods for the segmentation of degraded images of the iris. Markov chains have been well solicited to solve image segmentation problems. In this context, a feasibility study of unsupervised segmentation into regions of degraded iris images by Markov chains was performed. Different image transformations and different segmentation methods for parameters initialization have been studied and compared. Optimal modeling has been inserted in iris recognition system (with grayscale images) to produce a comparison with the existing methods. Finally, an extension of the modeling based on the hidden Markov chains has been developed in order to realize an unsupervised segmentation of the iris images acquired in visible light
Thivin, Solenne. „Détection automatique de cibles dans des fonds complexes. Pour des images ou séquences d'images“. Thesis, Université Paris-Saclay (ComUE), 2015. http://www.theses.fr/2015SACLS235/document.
Der volle Inhalt der QuelleDuring this PHD, we developped an detection algorithm. Our principal objective was to detect small targets in a complex background like clouds for example.For this, we used the spatial covariate structure of the real images.First, we developped a collection of models for this covariate structure. Then, we selected a special model in the previous collection. Once the model selected, we applied the likelihood ratio test to detect the potential targets.We finally studied the performances of our algorithm by testing it on simulated and real images
Negin, Farhood. „Vers une reconnaissance des activités humaines non supervisées et des gestes dans les vidéos“. Thesis, Université Côte d'Azur (ComUE), 2018. http://www.theses.fr/2018AZUR4246/document.
Der volle Inhalt der QuelleThe main goal of this thesis is to propose a complete framework for automatic discovery, modeling and recognition of human activities in videos. In order to model and recognize activities in long-term videos, we propose a framework that combines global and local perceptual information from the scene and accordingly constructs hierarchical activity models. In the first variation of the framework, a supervised classifier based on Fisher vector is trained and the predicted semantic labels are embedded in the constructed hierarchical models. In the second variation, to have a completely unsupervised framework, rather than embedding the semantic labels, the trained visual codebooks are stored in the models. Finally, we evaluate the proposed frameworks on two realistic Activities of Daily Living datasets recorded from patients in a hospital environment. Furthermore, to model fine motions of human body, we propose four different gesture recognition frameworks where each framework accepts one or combination of different data modalities as input. We evaluate the developed frameworks in the context of medical diagnostic test namely Praxis. Praxis test is a gesture-based diagnostic test, which has been accepted as a diagnostically indicative of cortical pathologies such as Alzheimer’s disease. We suggest a new challenge in gesture recognition, which is to obtain an objective opinion about correct and incorrect performances of very similar gestures. The experiments show effectiveness of our deep learning based approach in gesture recognition and performance assessment tasks
Cherdo, Yann. „Détection d'anomalie non supervisée sur les séries temporelle à faible coût énergétique utilisant les SNNs“. Electronic Thesis or Diss., Université Côte d'Azur, 2024. http://www.theses.fr/2024COAZ4018.
Der volle Inhalt der QuelleIn the context of the predictive maintenance of the car manufacturer Renault, this thesis aims at providing low-power solutions for unsupervised anomaly detection on time-series. With the recent evolution of cars, more and more data are produced and need to be processed by machine learning algorithms. This processing can be performed in the cloud or directly at the edge inside the car. In such a case, network bandwidth, cloud services costs, data privacy management and data loss can be saved. Embedding a machine learning model inside a car is challenging as it requires frugal models due to memory and processing constraints. To this aim, we study the usage of spiking neural networks (SNNs) for anomaly detection, prediction and classification on time-series. SNNs models' performance and energy costs are evaluated in an edge scenario using generic hardware models that consider all calculation and memory costs. To leverage as much as possible the sparsity of SNNs, we propose a model with trainable sparse connections that consumes half the energy compared to its non-sparse version. This model is evaluated on anomaly detection public benchmarks, a real use-case of anomaly detection from Renault Alpine cars, weather forecasts and the google speech command dataset. We also compare its performance with other existing SNN and non-spiking models. We conclude that, for some use-cases, spiking models can provide state-of-the-art performance while consuming 2 to 8 times less energy. Yet, further studies should be undertaken to evaluate these models once embedded in a car. Inspired by neuroscience, we argue that other bio-inspired properties such as attention, sparsity, hierarchy or neural assemblies dynamics could be exploited to even get better energy efficiency and performance with spiking models. Finally, we end this thesis with an essay dealing with cognitive neuroscience, philosophy and artificial intelligence. Diving into conceptual difficulties linked to consciousness and considering the deterministic mechanisms of memory, we argue that consciousness and the self could be constitutively independent from memory. The aim of this essay is to question the nature of humans by contrast with the ones of machines and AI
Germain, Mathieu. „L’estimation de distribution à l'aide d'un autoencodeur“. Mémoire, Université de Sherbrooke, 2015. http://hdl.handle.net/11143/6910.
Der volle Inhalt der QuelleSchreuder, Nicolas. „A study of some trade-offs in statistical learning : online learning, generative models and fairness“. Electronic Thesis or Diss., Institut polytechnique de Paris, 2021. http://www.theses.fr/2021IPPAG004.
Der volle Inhalt der QuelleMachine learning algorithms are celebrated for their impressive performance on many tasksthat we thought were dedicated to human minds, from handwritten digits recognition (LeCunet al. 1990) to cancer prognosis (Kourou et al. 2015). Nevertheless, as machine learning becomes more and more ubiquitous in our daily lives, there is a growing need for precisely understanding their behaviours and their limits.Statistical learning theory is the branch of machine learning which aims at providing a powerful modelling formalism for inference problems as well as a better understanding of the statistical properties of learning algorithms.Importantly, statistical learning theory allows one to (i) get a better understanding of the cases in which an algorithm performs well (ii) quantify trade-offs inherent to learning for better-informed algorithmic choices (iii) provide insights to develop new algorithms which will eventually outperform existing ones or tackle new tasks. Relying on the statistical learning framework, this thesis presents contributions related to three different learning problems: online learning, learning generative models and, finally, fair learning.In the online learning setup -- in which the sample size is not known in advance -- we provide general anytime deviation bounds (or confidence intervals) whose width has the rate given in the Law of Iterated Logarithm for a general class of convex M-estimators -- comprising the mean, the median, quantiles, Huber’s M-estimators.Regarding generative models, we propose a convenient framework for studying adversarial generative models (Goodfellow et al. 2014) from a statistical perspective to assess the impact of (eventual) low intrinsic dimensionality of the data on the error of the generative model. In our framework, we establish non-asymptotic risk bounds for the Empirical Risk Minimizer (ERM).Finally, our work on fair learning consists in a broad study of the Demographic Parity (DP) constraint, a popular constraint in the fair learning literature. DP essentially constrains predictors to treat groups defined by a sensitive attribute (e.g., gender or ethnicity) to be “treated the same”. In particular, we propose a statistical minimax framework to precisely quantify the cost in risk of introducing this constraint in the regression setting
Gal, Jocelyn. „Application d’algorithmes de machine learning pour l’exploitation de données omiques en oncologie“. Electronic Thesis or Diss., Université Côte d'Azur (ComUE), 2019. http://theses.univ-cotedazur.fr/2019AZUR6026.
Der volle Inhalt der QuelleThe development of computer science in medicine and biology has generated a large volume of data. The complexity and the amount of information to be integrated for optimal decision-making in medicine have largely exceeded human capacities. These data includes demographic, clinical and radiological variables, but also biological variables and particularly omics (genomics, proteomics, transcriptomics and metabolomics) characterized by a large number of measured variables relatively to a generally small number of patients. Their analysis represents a real challenge as they are frequently "noisy" and associated with situations of multi-colinearity. Nowadays, computational power makes it possible to identify clinically relevant models within these sets of data by using machine learning algorithms. Through this thesis, our goal is to apply supervised and unsupervised learning methods, to large biological data, in order to participate in the optimization of the classification and therapeutic management of patients with various types of cancer. In the first part of this work a supervised learning method is applied to germline immunogenetic data to predict the efficacy and toxicity of immune checkpoint inhibitor therapy. In the second part, different unsupervised learning methods are compared to evaluate the contribution of metabolomics in the diagnosis and management of breast cancer. Finally, the third part of this work aims to expose the contribution that simulated therapeutic trials can make in biomedical research. The application of machine learning methods in oncology offers new perspectives to clinicians allowing them to make diagnostics faster and more accurately, or to optimize therapeutic management in terms of efficacy and toxicity
Jouffroy, Emma. „Développement de modèles non supervisés pour l'obtention de représentations latentes interprétables d'images“. Electronic Thesis or Diss., Bordeaux, 2024. http://www.theses.fr/2024BORD0050.
Der volle Inhalt der QuelleThe Laser Megajoule (LMJ) is a large research device that simulates pressure and temperature conditions similar to those found in stars. During experiments, diagnostics are guided into an experimental chamber for precise positioning. To minimize the risks associated with human error in such an experimental context, the automation of an anti-collision system is envisaged. This involves the design of machine learning tools offering reliable decision levels based on the interpretation of images from cameras positioned in the chamber. Our research focuses on probabilistic generative neural methods, in particular variational auto-encoders (VAEs). The choice of this class of models is linked to the fact that it potentially enables access to a latent space directly linked to the properties of the objects making up the observed scene. The major challenge is to study the design of deep network models that effectively enable access to such a fully informative and interpretable representation, with a view to system reliability. The probabilistic formalism intrinsic to VAE allows us, if we can trace back to such a representation, to access an analysis of the uncertainties of the encoded information
Jabiri, Fouad. „Applications de méthodes de classification non supervisées à la détection d'anomalies“. Master's thesis, Université Laval, 2020. http://hdl.handle.net/20.500.11794/67914.
Der volle Inhalt der QuelleIn this thesis, we will first present the binary tree partitioning algorithm and isolation forests. Binary trees are very popular classifiers in supervised machine learning. The isolation forest belongs to the family of unsupervised methods. It is an ensemble of binary trees used in common to isolate outlying instances. Subsequently, we will present the approach that we have named "Exponential smoothig" (or "pooling"). This technique consists in encoding sequences of variables of different lengths into a single vector of fixed size. Indeed, the objective of this thesis is to apply the algorithm of isolation forests to identify anomalies in insurance claim forms available in the database of a large Canadian insurance company in order to detect cases of fraud. However, a form is a sequence of claims. Each claim is characterized by a set of variables and thus it will be impossible to apply the isolation forest algorithm directly to this kind of data. It is for this reason that we are going to apply Exponential smoothing. Our application effectively isolates claims and abnormal forms, and we find that the latter tend to be audited by the company more often than regular forms.
Pellegrini, Thomas. „Transcription automatique de langues peu dotées“. Phd thesis, Université Paris Sud - Paris XI, 2008. http://tel.archives-ouvertes.fr/tel-00619657.
Der volle Inhalt der QuelleWashha, Mahdi. „Information quality in online social media and big data collection : an example of Twitter spam detection“. Thesis, Toulouse 3, 2018. http://www.theses.fr/2018TOU30080/document.
Der volle Inhalt der QuelleThe popularity of OSM is mainly conditioned by the integrity and the quality of UGC as well as the protection of users' privacy. Based on the definition of information quality as fitness for use, the high usability and accessibility of OSM have exposed many information quality (IQ) problems which consequently decrease the performance of OSM dependent applications. Such problems are caused by ill-intentioned individuals who misuse OSM services to spread different kinds of noisy information, including fake information, illegal commercial content, drug sales, mal- ware downloads, and phishing links. The propagation and spreading of noisy information cause enormous drawbacks related to resources consumptions, decreasing quality of service of OSM-based applications, and spending human efforts. The majority of popular social networks (e.g., Facebook, Twitter, etc) over the Web 2.0 is daily attacked by an enormous number of ill-intentioned users. However, those popular social networks are ineffective in handling the noisy information, requiring several weeks or months to detect them. Moreover, different challenges stand in front of building a complete OSM-based noisy information filtering methods that can overcome the shortcomings of OSM information filters. These challenges are summarized in: (i) big data; (ii) privacy and security; (iii) structure heterogeneity; (iv) UGC format diversity; (v) subjectivity and objectivity; (vi) and service limitations In this thesis, we focus on increasing the quality of social UGC that are published and publicly accessible in forms of posts and profiles over OSNs through addressing in-depth the stated serious challenges. As the social spam is the most common IQ problem appearing over the OSM, we introduce a design of two generic approaches for detecting and filtering out the spam content. The first approach is for detecting the spam posts (e.g., spam tweets) in a real-time stream, while the other approach is dedicated for handling a big data collection of social profiles (e.g., Twitter accounts)
Frery, Jordan. „Ensemble Learning for Extremely Imbalced Data Flows“. Thesis, Lyon, 2019. http://www.theses.fr/2019LYSES034.
Der volle Inhalt der QuelleMachine learning is the study of designing algorithms that learn from trainingdata to achieve a specific task. The resulting model is then used to predict overnew (unseen) data points without any outside help. This data can be of manyforms such as images (matrix of pixels), signals (sounds,...), transactions (age,amount, merchant,...), logs (time, alerts, ...). Datasets may be defined to addressa specific task such as object recognition, voice identification, anomaly detection,etc. In these tasks, the knowledge of the expected outputs encourages a supervisedlearning approach where every single observed data is assigned to a label thatdefines what the model predictions should be. For example, in object recognition,an image could be associated with the label "car" which suggests that the learningalgorithm has to learn that a car is contained in this picture, somewhere. This is incontrast with unsupervised learning where the task at hand does not have explicitlabels. For example, one popular topic in unsupervised learning is to discoverunderlying structures contained in visual data (images) such as geometric formsof objects, lines, depth, before learning a specific task. This kind of learning isobviously much harder as there might be potentially an infinite number of conceptsto grasp in the data. In this thesis, we focus on a specific scenario of thesupervised learning setting: 1) the label of interest is under represented (e.g.anomalies) and 2) the dataset increases with time as we receive data from real-lifeevents (e.g. credit card transactions). In fact, these settings are very common inthe industrial domain in which this thesis takes place
Félicien, Vallet. „Structuration automatique de talk shows télévisés“. Phd thesis, Télécom ParisTech, 2011. http://pastel.archives-ouvertes.fr/pastel-00635495.
Der volle Inhalt der QuelleVelcin, Julien. „Extraction automatique de stéréotypes à partir de données symboliques et lacunaires“. Paris 6, 2005. http://www.theses.fr/2005PA066465.
Der volle Inhalt der QuelleLiu, Jingshu. „Unsupervised cross-lingual representation modeling for variable length phrases“. Thesis, Nantes, 2020. http://www.theses.fr/2020NANT4009.
Der volle Inhalt der QuelleSignificant advances have been achieved in bilingual word-level alignment from comparable corpora, yet the challenge remains for phrase-level alignment. Traditional methods to phrase alignment can only handle phrase of equal length, while word embedding based approaches learn phrase embeddings as individual vocabulary entries suffer from the data sparsity and cannot handle out of vocabulary phrases. Since bilingual alignment is a vector comparison task, phrase representation plays a key role. In this thesis, we study the approaches for unified phrase modeling and cross-lingual phrase alignment, ranging from co-occurrence models to most recent neural state-of-the-art approaches. We review supervised and unsupervised frameworks for modeling cross-lingual phrase representations. Two contributions are proposed in this work. First, a new architecture called tree-free recursive neural network (TF-RNN) for modeling phrases of variable length which, combined with a wrapped context prediction training objective, outperforms the state-of-the-art approaches on monolingual phrase synonymy task with only plain text training data. Second, for cross-lingual modeling, we propose to incorporate an architecture derived from TF-RNN in an encoder-decoder model with a pseudo back translation mechanism inspired by unsupervised neural machine translation. Our proposition improves significantly bilingual alignment of different length phrases
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
Khaleghi, Azadeh. „Sur quelques problèmes non-supervisés impliquant des séries temporelles hautement dépendantes“. Phd thesis, Université des Sciences et Technologie de Lille - Lille I, 2013. http://tel.archives-ouvertes.fr/tel-00920333.
Der volle Inhalt der QuellePantin, Jérémie. „Détection et caractérisation sémantique de données textuelles aberrantes“. Electronic Thesis or Diss., Sorbonne université, 2023. https://accesdistant.sorbonne-universite.fr/login?url=https://theses-intra.sorbonne-universite.fr/2023SORUS347.pdf.
Der volle Inhalt der QuelleMachine learning answers to the problem of handling dedicated tasks with a wide variety of data. Such algorithms can be either simple or difficult to handle depending of the data. Low dimensional data (2-dimension or 3-dimension) with an intuitive representation (average of baguette price by years) are easier to interpret/explain for a human than data with thousands of dimensions. For low dimensional data, the error leads to a significant shift against normal data, but for the case of high dimensional data it is different. Outlier detection (or anomaly detection, or novelty detection) is the study of outlying observations for detecting what is normal and abnormal. Methods that perform such task are algorithms, methods or models that are based on data distributions. Different families of approaches can be found in the literature of outlier detection, and they are mainly independent of ground truth. They perform outlier analysis by detecting the principal behaviors of majority of observations. Thus, data that differ from normal distribution are considered noise or outlier. We detail the application of outlier detection with text. Despite recent progress in natural language processing, computer still lack profound understanding of human language in absence of information. For instance, the sentence "A smile is a curve that set everything straight" has several levels of understanding and a machine can encounter hardship to chose the right level of lecture. This thesis presents the analysis of high-dimensional outliers, applied to text. Recent advances in anomaly detection and outlier detection are not significantly represented with text data and we propose to highlight the main differences with high-dimensional outliers. We also approach ensemble methods that are nearly nonexistent in the literature for our context. Finally, an application of outlier detection for elevate results on abstractive summarization is conducted. We propose GenTO, a method that prepares and generates split of data in which anomalies and outliers are inserted. Based on this method, evaluation and benchmark of outlier detection approaches is proposed with documents. The proposed taxonomy allow to identify difficult and hierarchised outliers that the literature tackles without knowing. Also, learning without supervision often leads models to rely in some hyperparameter. For instance, Local Outlier Factor relies to the k-nearest neighbors for computing the local density. Thus, choosing the right value for k is crucial. In this regard, we explore the influence of such parameter for text data. While choosing one model can leads to obvious bias against real-world data, ensemble methods allow to mitigate such problem. They are particularly efficient with outlier analysis. Indeed, the selection of several values for one hyperparameter can help to detect strong outliers.Importance is then tackled and can help a human to understand the output of black box model. Thus, the interpretability of outlier detection models is questioned. We find that for numerous dataset, a low number of features can be selected as oracle. The association of complete models and restrained models helps to mitigate the black-box effect of some approaches. In some cases, outlier detection refers to noise removal or anomaly detection. Some applications can benefit from the characteristic of such task. Mail spam detection and fake news detection are one example, but we propose to use outlier detection approaches for weak signal exploration in marketing project. Thus, we find that the model of the literature help to improve unsupervised abstractive summarization, and also to find weak signals in text
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 QuelleBailleux, Caroline. „Métabolomique du cancer du sein localisé à haut risque de récidive“. Electronic Thesis or Diss., Université Côte d'Azur, 2023. http://www.theses.fr/2023COAZ6017.
Der volle Inhalt der QuelleBreast cancer is a heterogeneous disease with multiple histological, biological, and molecular subtypes. Several fundamental studies have highlighted the activation of specific metabolic pathways in aggressive breast cancers. The aim of this thesis was to identify a signature or markers of the metabolome in localized breast cancer at high risk of recurrence.Our initial studies were based on the retrospective inclusion of 52 patients with localized breast cancer treated at the Antoine Lacassagne Center in Nice. We also analyzed diagnostic biopsies from a cohort of 49 patients treated with neo-adjuvant chemotherapy at the Georges-François Leclerc Center in Dijon for locally advanced breast cancer. After extraction, separation, and concentration of metabolites from diagnostic biopsies and resected tumors, we performed metabolomic profiling using LC-MS/MS to identify and quantify metabolites relatively, followed by biological and statistical analysis.First, we compared the performance of 5 unsupervised machine learning methods (PCA k-means, sparse k-means, spectral clustering, SIMLR, and k-sparse) to identify groups of breast cancer patients. This analysis was only performed on the cohort from Nice.In Article 1, the clusters obtained using the 5 unsupervised machine learning methods were compared. The five methods identified three groups of patients, distinguished by their supposed prognosis (favorable group 1, intermediate group 2, unfavorable group 3), with different clinical and biological profiles. The SIMLR and K-sparse methods were the most effective in terms of clustering. The most discriminating metabolic pathways were glycolysis, glutaminolysis, and amino acid metabolism. The simulated "in-silico" survival analysis (PREDICT tool) revealed a significant difference between the 3 groups for 5-year and 10-year specific survival.In Article 2, survival analyses were performed based on actual patient survival data. Each patient was assigned to his prognostic group established by the 5 unsupervised machine learning methods. Groups 1 and 2 were combined and compared to group 3. The median follow-up was extended to 85.8 months. Bootstrap optimization was applied. The PCA k-means, K-sparse, and Spectral clustering methods achieved the best results for predicting 2-year progression-free survival. The PCA k-means method had the best performance. However, CSS and OS analyses revealed discrepancies between the 5 unsupervised machine learning methods.Simultaneously, a supervised analysis comparing high-grade tumors to low/intermediate grade tumors was conducted to determine the metabolites involved in tumor aggressiveness (Article 3). The Nice cohort was used as a training cohort, while the Dijon cohort was used for external validation. The metabolomic signature was composed of 12 metabolites. The AUCs for the training and validation cohorts were greater than 0.88. Thus, the model could distinguish high-grade tumors from low/intermediate grade tumors with a probability of nearly 90%. We identified several biomarkers of tumor aggressiveness, such as N1, N12 diacetylspermine and tryptophan catabolites (kynurenine and serotonin), which are involved in inhibiting the immune response.These studies open up new perspectives on the underlying biological mechanisms of tumor aggressiveness. Furthermore, the identified biomarkers will allow the development of new strategies. However, analyses on larger populations are necessary
Franceschi, Jean-Yves. „Apprentissage de représentations et modèles génératifs profonds dans les systèmes dynamiques“. Electronic Thesis or Diss., Sorbonne université, 2022. http://www.theses.fr/2022SORUS014.
Der volle Inhalt der QuelleThe recent rise of deep learning has been motivated by numerous scientific breakthroughs, particularly regarding representation learning and generative modeling. However, most of these achievements have been obtained on image or text data, whose evolution through time remains challenging for existing methods. Given their importance for autonomous systems to adapt in a constantly evolving environment, these challenges have been actively investigated in a growing body of work. In this thesis, we follow this line of work and study several aspects of temporality and dynamical systems in deep unsupervised representation learning and generative modeling. Firstly, we present a general-purpose deep unsupervised representation learning method for time series tackling scalability and adaptivity issues arising in practical applications. We then further study in a second part representation learning for sequences by focusing on structured and stochastic spatiotemporal data: videos and physical phenomena. We show in this context that performant temporal generative prediction models help to uncover meaningful and disentangled representations, and conversely. We highlight to this end the crucial role of differential equations in the modeling and embedding of these natural sequences within sequential generative models. Finally, we more broadly analyze in a third part a popular class of generative models, generative adversarial networks, under the scope of dynamical systems. We study the evolution of the involved neural networks with respect to their training time by describing it with a differential equation, allowing us to gain a novel understanding of this generative model
Buhot, Arnaud. „Etude de propriétés d'apprentissage supervisé et non supervisé par des méthodes de Physique Statistique“. Phd thesis, Université Joseph Fourier (Grenoble), 1999. http://tel.archives-ouvertes.fr/tel-00001642.
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
Tiomoko, ali Hafiz. „Nouvelles méthodes pour l’apprentissage non-supervisé en grandes dimensions“. Thesis, Université Paris-Saclay (ComUE), 2018. http://www.theses.fr/2018SACLC074/document.
Der volle Inhalt der QuelleSpurred by recent advances on the theoretical analysis of the performances of the data-driven machine learning algorithms, this thesis tackles the performance analysis and improvement of high dimensional data and graph clustering. Specifically, in the first bigger part of the thesis, using advanced tools from random matrix theory, the performance analysis of spectral methods on dense realistic graph models and on high dimensional kernel random matrices is performed through the study of the eigenvalues and eigenvectors of the similarity matrices characterizing those data. New improved methods are proposed and are shown to outperform state-of-the-art approaches. In a second part, a new algorithm is proposed for the detection of heterogeneous communities from multi-layer graphs using variational Bayes approaches to approximate the posterior distribution of the sought variables. The proposed methods are successfully applied to synthetic benchmarks as well as real-world datasets and are shown to outperform standard approaches to clustering in those specific contexts
Boniol, Paul. „Detection of anomalies and identification of their precursors in large data series collections“. Electronic Thesis or Diss., Université Paris Cité, 2021. http://www.theses.fr/2021UNIP5206.
Der volle Inhalt der QuelleExtensive collections of data series are becoming a reality in a large number of scientific and social domains. There is, therefore, a growing interest and need to elaborate efficient techniques to analyze and process these data, such as in finance, environmental sciences, astrophysics, neurosciences, engineering. Informally, a data series is an ordered sequence of points or values. Once these series are collected and available, users often need to query them. These queries can be simple, such as the selection of time interval, but also complex, such as the similarities search or the detection of anomalies, often synonymous with malfunctioning of the system under study, or sudden and unusual evolution likely undesired. This last type of analysis represents a crucial problem for applications in a wide range of domains, all sharing the same objective: to detect anomalies as soon as possible to avoid critical events. Therefore, in this thesis, we address the following three objectives: (i) retrospective unsupervised subsequence anomaly detection in data series. (ii) unsupervised detection of anomalies in data streams. (iii) classification explanation of known anomalies in data series in order to identify possible precursors. This manuscript first presents the industrial context that motivated this thesis, fundamental definitions, a taxonomy of data series, and state-of-the-art anomaly detection methods. We then present our contributions along the three axes mentioned above. First, we describe two original solutions, NormA (that aims to build a weighted set of subsequences that represent the different behaviors of the data series) and Series2Graph (that transform the data series in a directed graph), for the task of unsupervised detection of anomalous subsequences in static data series. Secondly, we present the SAND (inspired from NormA) method for unsupervised detection of anomalous subsequences in data streams. Thirdly, we address the problem of the supervised identification of precursors. We subdivide this task into two generic problems: the supervised classification of time series and the explanation of this classification’s results by identifying discriminative subsequences. Finally, we illustrate the applicability and interest of our developments through an application concerning the identification of undesirable vibration precursors occurring in water supply pumps in the French nuclear power plants of EDF
Courjault-Rade, Vincent. „Ballstering : un algorithme de clustering dédié à de grands échantillons“. Thesis, Toulouse 3, 2018. http://www.theses.fr/2018TOU30126/document.
Der volle Inhalt der QuelleBallstering belongs to the machine learning methods that aim to group in classes a set of objects that form the studied dataset, without any knowledge of true classes within it. This type of methods, of which k-means is one of the most famous representative, are named clustering methods. Recently, a new clustering algorithm "Fast Density Peak Clustering" (FDPC) has aroused great interest from the scientific community for its innovating aspect and its efficiency on non-concentric distributions. However this algorithm showed a such complexity that it can't be applied with ease on large datasets. Moreover, we have identified several weaknesses that impact the quality results and the presence of a general parameter dc difficult to choose while having a significant impact on the results. In view of those limitations, we reworked the principal idea of FDPC in a new light and modified it successively to finally create a distinct algorithm that we called Ballstering. The work carried out during those three years can be summarised by the conception of this clustering algorithm especially designed to be effective on large datasets. As its Precursor, Ballstering works in two phases: An estimation density phase followed by a clustering step. Its conception is mainly based on a procedure that handle the first step with a lower complexity while avoiding at the same time the difficult choice of dc, which becomes automatically defined according to local density. We name ICMDW this procedure which represent a consistent part of our contributions. We also overhauled cores definitions of FDPC and entirely reworked the second phase (relying on the graph structure of ICMDW's intermediate results), to finally produce an algorithm that overcome all the limitations that we have identified
Kannan, Hariprasad. „Quelques applications de l’optimisation numérique aux problèmes d’inférence et d’apprentissage“. Thesis, Université Paris-Saclay (ComUE), 2018. http://www.theses.fr/2018SACLC067/document.
Der volle Inhalt der QuelleNumerical optimization and machine learning have had a fruitful relationship, from the perspective of both theory and application. In this thesis, we present an application oriented take on some inference and learning problems. Linear programming relaxations are central to maximum a posteriori (MAP) inference in discrete Markov Random Fields (MRFs). Especially, inference in higher-order MRFs presents challenges in terms of efficiency, scalability and solution quality. In this thesis, we study the benefit of using Newton methods to efficiently optimize the Lagrangian dual of a smooth version of the problem. We investigate their ability to achieve superior convergence behavior and to better handle the ill-conditioned nature of the formulation, as compared to first order methods. We show that it is indeed possible to obtain an efficient trust region Newton method, which uses the true Hessian, for a broad range of MAP inference problems. Given the specific opportunities and challenges in the MAP inference formulation, we present details concerning (i) efficient computation of the Hessian and Hessian-vector products, (ii) a strategy to damp the Newton step that aids efficient and correct optimization, (iii) steps to improve the efficiency of the conjugate gradient method through a truncation rule and a pre-conditioner. We also demonstrate through numerical experiments how a quasi-Newton method could be a good choice for MAP inference in large graphs. MAP inference based on a smooth formulation, could greatly benefit from efficient sum-product computation, which is required for computing the gradient and the Hessian. We show a way to perform sum-product computation for trees with sparse clique potentials. This result could be readily used by other algorithms, also. We show results demonstrating the usefulness of our approach using higher-order MRFs. Then, we discuss potential research topics regarding tightening the LP relaxation and parallel algorithms for MAP inference.Unsupervised learning is an important topic in machine learning and it could potentially help high dimensional problems like inference in graphical models. We show a general framework for unsupervised learning based on optimal transport and sparse regularization. Optimal transport presents interesting challenges from an optimization point of view with its simplex constraints on the rows and columns of the transport plan. We show one way to formulate efficient optimization problems inspired by optimal transport. This could be done by imposing only one set of the simplex constraints and by imposing structure on the transport plan through sparse regularization. We show how unsupervised learning algorithms like exemplar clustering, center based clustering and kernel PCA could fit into this framework based on different forms of regularization. We especially demonstrate a promising approach to address the pre-image problem in kernel PCA. Several methods have been proposed over the years, which generally assume certain types of kernels or have too many hyper-parameters or make restrictive approximations of the underlying geometry. We present a more general method, with only one hyper-parameter to tune and with some interesting geometric properties. From an optimization point of view, we show how to compute the gradient of a smooth version of the Schatten p-norm and how it can be used within a majorization-minimization scheme. Finally, we present results from our various experiments
Debard, Quentin. „Automatic learning of next generation human-computer interactions“. Thesis, Lyon, 2020. http://www.theses.fr/2020LYSEI036.
Der volle Inhalt der QuelleArtificial Intelligence (AI) and Human-Computer Interactions (HCIs) are two research fields with relatively few common work. HCI specialists usually design the way we interact with devices directly from observations and measures of human feedback, manually optimizing the user interface to better fit users’ expectations. This process is hard to optimize: ergonomy, intuitivity and ease of use are key features in a User Interface (UI) that are too complex to be simply modelled from interaction data. This drastically restrains the possible uses of Machine Learning (ML) in this design process. Currently, ML in HCI is mostly applied to gesture recognition and automatic display, e.g. advertisement or item suggestion. It is also used to fine tune an existing UI to better optimize it, but as of now it does not participate in designing new ways to interact with computers. Our main focus in this thesis is to use ML to develop new design strategies for overall better UIs. We want to use ML to build intelligent – understand precise, intuitive and adaptive – user interfaces using minimal handcrafting. We propose a novel approach to UI design: instead of letting the user adapt to the interface, we want the interface and the user to adapt mutually to each other. The goal is to reduce human bias in protocol definition while building co-adaptive interfaces able to further fit individual preferences. In order to do so, we will put to use the different mechanisms available in ML to automatically learn behaviors, build representations and take decisions. We will be experimenting on touch interfaces, as these interfaces are vastly used and can provide easily interpretable problems. The very first part of our work will focus on processing touch data and use supervised learning to build accurate classifiers of touch gestures. The second part will detail how Reinforcement Learning (RL) can be used to model and learn interaction protocols given user actions. Lastly, we will combine these RL models with unsupervised learning to build a setup allowing for the design of new interaction protocols without the need for real user data
Gal, Viviane. „Vers une nouvelle Interaction Homme Environnement dans les jeux vidéo et pervasifs : rétroaction biologique et états émotionnels : apprentissage profond non supervisé au service de l'affectique“. Electronic Thesis or Diss., Paris, CNAM, 2019. http://www.theses.fr/2019CNAM1269.
Der volle Inhalt der QuelleLiving exceptional moments, experiencing thrills, well-being, blooming, are often part of our dreams or aspirations. We choose various ways to get there like games. Whether the player is looking for originality, challenges, discovery, a story, or other goals, emotional states are the purpose of his quest. He remains until the game gives him pleasure, sensations. How bring them there? We are developing a new human environment interaction that takes into account and adapts to emotions. We address video or pervasive games or other applications. Through this goal, players should not be bothered by interfaces, or biosensors invasivness. This work raises two questions:- Can we discover emotional states based on physiological measurements from contact biosensors?- If so, can these sensors be replaced by remote, non-invasive devices and produce the same results?The models we have developed propose solutions based on unsupervised machine learning methods. We also present remote measurements technics and explain our future works in a new field we call affectics
Oquab, Maxime. „Convolutional neural networks : towards less supervision for visual recognition“. Thesis, Paris Sciences et Lettres (ComUE), 2018. http://www.theses.fr/2018PSLEE061.
Der volle Inhalt der QuelleConvolutional Neural Networks are flexible learning algorithms for computer vision that scale particularly well with the amount of data that is provided for training them. Although these methods had successful applications already in the ’90s, they were not used in visual recognition pipelines because of their lesser performance on realistic natural images. It is only after the amount of data and the computational power both reached a critical point that these algorithms revealed their potential during the ImageNet challenge of 2012, leading to a paradigm shift in visual recogntion. The first contribution of this thesis is a transfer learning setup with a Convolutional Neural Network for image classification. Using a pre-training procedure, we show that image representations learned in a network generalize to other recognition tasks, and their performance scales up with the amount of data used in pre-training. The second contribution of this thesis is a weakly supervised setup for image classification that can predict the location of objects in complex cluttered scenes, based on a dataset indicating only with the presence or absence of objects in training images. The third contribution of this thesis aims at finding possible paths for progress in unsupervised learning with neural networks. We study the recent trend of Generative Adversarial Networks and propose two-sample tests for evaluating models. We investigate possible links with concepts related to causality, and propose a two-sample test method for the task of causal discovery. Finally, building on a recent connection with optimal transport, we investigate what these generative algorithms are learning from unlabeled data
Le, Lan Gaël. „Analyse en locuteurs de collections de documents multimédia“. Thesis, Le Mans, 2017. http://www.theses.fr/2017LEMA1020/document.
Der volle Inhalt der QuelleThe task of speaker diarization and linking aims at answering the question "who speaks and when?" in a collection of multimedia recordings. It is an essential step to index audiovisual contents. The task of speaker diarization and linking firstly consists in segmenting each recording in terms of speakers, before linking them across the collection. Aim is, to identify each speaker with a unique anonymous label, even for speakers appearing in multiple recordings, without any knowledge of their identity or number. The challenge of the cross-recording linking is the modeling of the within-speaker/across-recording variability: depending on the recording, a same speaker can appear in multiple acoustic conditions (in a studio, in the street...). The thesis proposes two methods to overcome this issue. Firstly, a novel neural variability compensation method is proposed, using the triplet-loss paradigm for training. Secondly, an iterative unsupervised domain adaptation process is presented, in which the system exploits the information (even inaccurate) about the data it processes, to enhance its performances on the target acoustic domain. Moreover, novel ways of analyzing the results in terms of speaker are explored, to understand the actual performance of a diarization and linking system, beyond the well-known Diarization Error Rate (DER). Systems and methods are evaluated on two TV shows of about 40 episodes, using either a global, or longitudinal linking architecture, and state of the art speaker modeling (i-vector)