In today’s machine learning research, probabilistic graphical models are used extensively to model complicated systems with uncertainty, to help understanding of the problems, and to help inference and predict unknown events. For inference tasks, exact inference methods such as junction tree algorithms exist, but they suffer from exponential growth of cluster size and thus is not able to handle large and highly connected graphs. Approximate inference methods do not try to find exact probabilities, but rather give results that improve as algorithm runs. Gibbs sampling, as one of the approximate inference methods, has gained lots of traction and is used extensively in inference tasks, due to its ease of understanding and implementation. However, as problem size grows, even the faster algorithm needs a speed boost to meet application requirement. The number of variables in an application graphical model can range from tens of thousands to billions, depending on problem domain. The original sequential Gibbs sampling may not return satisfactory result in limited time. Thus, in this thesis, we investigate in ways to speed up Gibbs sampling. We will study ways to do better initialization, blocking variables to be sampled together, as well as using simulated annealing. These are the methods that modifies the algorithm itself. We will also investigate in ways to parallelize the algorithm. An algorithm is parallelizable if some steps do not depend on other steps, and we will find out such dependency in Gibbs sampling. We will discuss how the choice of different hardware and software architecture will affect the parallelization result. We will use optical flow problem as an example to demonstrate the various speed up methods we investigated. An optical flow method tries to find out the movements of small image patches between two images in a temporal sequence. We demonstrate how we can model it using probabilistic graphical model, and solve it using Gibbs sampling. The result of using sequential Gibbs sampling is demonstrated, with comparisons from using various speed up methods and other optical flow methods.

Objective: This thesis focuses on design and evaluation of a disease prediction system that be able to detect hidden and upcoming diseases of an individual. Unlike previous works that has typically relied on precise medical examinations to extract symptoms and risk factors for computing probability of occurrence of a disease, the proposed disease prediction system is based on similar patterns of disease comorbidity in population and the individual to evaluate the risk of a disease. Methods: We combine three machine learning algorithms to construct the prediction system: an item based recommendation system, a Bayesian graphical model and a rule based recommender. We also propose multiple similarity measures for the recommendation system, each useful in a particular condition. We finally show how best values of parameters of the system can be derived from optimization of cost function and ROC curve. Results: A permutation test is designed to evaluate accuracy of the prediction system accurately. Results showed considerable advantage of the proposed system in compare to an item based recommendation system and improvements of prediction if system is trained for each specific gender and race. Conclusion: The proposed system has been shown to be a competent method in accurately identifying potential diseases in patients with multiple diseases, just based on their disease records. The procedure also contains novel soft computing and machine learning ideas that can be used in prediction problems. The proposed system has the possibility of using more complex datasets that include timeline of diseases, disease networks and social network. This makes it an even more capable platform for disease prediction. Hence, this thesis contributes to improvement of the disease prediction field. / Graduate / 0800 / 0766 / 0984 / mehrdadmansouri@yahoo.com

A Factor Graph (FG -- http://en.wikipedia.org/wiki/Factor_graph) is a structure used to find solutions to problems that can be represented as a Probabilistic Graphical Model (PGM). They consist of interconnected variable nodes and factor nodes, which iteratively compute and pass messages to each other. FGs can be applied to solve decoding of forward error correcting codes, Markov chains and Markov Random Fields, Kalman Filtering, Fourier Transforms, and even some games such as Sudoku. In this paper, a framework is presented for rapid prototyping of hardware implementations of FG-based applications. The FG developer specifies aspects of the application, such as graphical structure, factor computation, and message passing algorithm, and the framework returns a design. A system of Python scripts and Verilog Hardware Description Language templates together are used to generate the HDL source code for the application. The generated designs are vendor/platform agnostic, but currently target the Xilinx Virtex-6-based ML605. The framework has so far been primarily applied to construct Low Density Parity Check (LDPC) decoders. The characteristics of a large basket of generated LDPC decoders, including contemporary 802.11n decoders, have been examined as a verification of the system and as a demonstration of its capabilities. As a further demonstration, the framework has been applied to construct a Sudoku solver. / Master of Science

Dans cette thèse, nous nous intéressons au problème spécifique de l'apprentissage de structure de modèles graphiques probabilistes, c'est-à-dire trouver la structure la plus efficace pour représenter une distribution, à partir seulement d'un ensemble d'échantillons D ∼ p(v). Dans une première partie, nous passons en revue les principaux modèles graphiques probabilistes de la littérature, des plus classiques (modèles dirigés, non-dirigés) aux plus avancés (modèles mixtes, cycliques etc.). Puis nous étudions particulièrement le problème d'apprentissage de structure de modèles dirigés (réseaux Bayésiens), et proposons une nouvelle méthode hybride pour l'apprentissage de structure, H2PC (Hybrid Hybrid Parents and Children), mêlant une approche à base de contraintes (tests statistiques d'indépendance) et une approche à base de score (probabilité postérieure de la structure). Dans un second temps, nous étudions le problème de la classification multi-label, visant à prédire un ensemble de catégories (vecteur binaire y P (0, 1)m) pour un objet (vecteur x P Rd). Dans ce contexte, l'utilisation de modèles graphiques probabilistes pour représenter la distribution conditionnelle des catégories prend tout son sens, particulièrement dans le but minimiser une fonction coût complexe. Nous passons en revue les principales approches utilisant un modèle graphique probabiliste pour la classification multi-label (Probabilistic Classifier Chain, Conditional Dependency Network, Bayesian Network Classifier, Conditional Random Field, Sum-Product Network), puis nous proposons une approche générique visant à identifier une factorisation de p(y|x) en distributions marginales disjointes, en s'inspirant des méthodes d'apprentissage de structure à base de contraintes. Nous démontrons plusieurs résultats théoriques, notamment l'unicité d'une décomposition minimale, ainsi que trois procédures quadratiques sous diverses hypothèses à propos de la distribution jointe p(x, y). Enfin, nous mettons en pratique ces résultats afin d'améliorer la classification multi-label avec les fonctions coût F-loss et zero-one loss / In this thesis, we address the specific problem of probabilistic graphical model structure learning, that is, finding the most efficient structure to represent a probability distribution, given only a sample set D ∼ p(v). In the first part, we review the main families of probabilistic graphical models from the literature, from the most common (directed, undirected) to the most advanced ones (chained, mixed etc.). Then we study particularly the problem of learning the structure of directed graphs (Bayesian networks), and we propose a new hybrid structure learning method, H2PC (Hybrid Hybrid Parents and Children), which combines a constraint-based approach (statistical independence tests) with a score-based approach (posterior probability of the structure). In the second part, we address the multi-label classification problem, which aims at assigning a set of categories (binary vector y P (0, 1)m) to a given object (vector x P Rd). In this context, probabilistic graphical models provide convenient means of encoding p(y|x), particularly for the purpose of minimizing general loss functions. We review the main approaches based on PGMs for multi-label classification (Probabilistic Classifier Chain, Conditional Dependency Network, Bayesian Network Classifier, Conditional Random Field, Sum-Product Network), and propose a generic approach inspired from constraint-based structure learning methods to identify the unique partition of the label set into irreducible label factors (ILFs), that is, the irreducible factorization of p(y|x) into disjoint marginal distributions. We establish several theoretical results to characterize the ILFs based on the compositional graphoid axioms, and obtain three generic procedures under various assumptions about the conditional independence properties of the joint distribution p(x, y). Our conclusions are supported by carefully designed multi-label classification experiments, under the F-loss and the zero-one loss functions

The aim of distributional semantics is to design computational techniques that can automatically learn the meanings of words from a body of text. The twin challenges are: how do we represent meaning, and how do we learn these representations? The current state of the art is to represent meanings as vectors - but vectors do not correspond to any traditional notion of meaning. In particular, there is no way to talk about 'truth', a crucial concept in logic and formal semantics. In this thesis, I develop a framework for distributional semantics which answers this challenge. The meaning of a word is not represented as a vector, but as a 'function', mapping entities (objects in the world) to probabilities of truth (the probability that the word is true of the entity). Such a function can be interpreted both in the machine learning sense of a classifier, and in the formal semantic sense of a truth-conditional function. This simultaneously allows both the use of machine learning techniques to exploit large datasets, and also the use of formal semantic techniques to manipulate the learnt representations. I define a probabilistic graphical model, which incorporates a probabilistic generalisation of model theory (allowing a strong connection with formal semantics), and which generates semantic dependency graphs (allowing it to be trained on a corpus). This graphical model provides a natural way to model logical inference, semantic composition, and context-dependent meanings, where Bayesian inference plays a crucial role. I demonstrate the feasibility of this approach by training a model on WikiWoods, a parsed version of the English Wikipedia, and evaluating it on three tasks. The results indicate that the model can learn information not captured by vector space models.

L'analyse de fiabilité fait partie intégrante de la conception et du fonctionnement du système, en particulier pour les systèmes exécutant des applications critiques. Des travaux récents ont montré l'intérêt d'utiliser les réseaux bayésiens dans le domaine de la fiabilité, pour modélisation la dégradation d'un système. Les modèles graphiques de durée sont un cas particulier des réseaux bayésiens, qui permettent de s'affranchir de la propriété markovienne des réseaux bayésiens dynamiques. Ils s'adaptent aux systèmes dont le temps de séjour dans chaque état n'est pas nécessairement distribué exponentiellement, comme c'est le cas dans la plupart des applications industrielles. Des travaux antérieurs ont toutefois montré des limitations à ces modèles en terme de capacité de stockage et de temps de calcul, en raison du caractère discret de la variable temps de séjour. Une solution pourrait consister à considérer une variable de durée continue. Selon les avis d'experts, les variables de temps de séjour suivent une distribution de Weibull dans de nombreux systèmes. L'objectif de la thèse est d'intégrer des variables de temps de séjour suivant une distribution de Weibull dans un modèle de durée graphique en proposant une nouvelle approche. Après une présentation des réseaux bayésiens, et plus particulièrement des modèles graphiques de durée et leur limitation, ce rapport s'attache à présenter le nouveau modèle permettant la modélisation du processus de dégradation. Ce nouveau modèle est appelé modèle graphique de durée hybride Weibull. Un algorithme original permettant l'inférence dans un tel réseau a été mis en place. L'étape suivante a été la validation de l'approche. Ne disposant pas de données, il a été nécessaire de simuler des séquences d'états du système. Différentes bases de données ainsi construites ont permis d'apprendre d'un part un modèle graphique de durée, et d'autre part un modèle graphique de durée hybride-Weibull, afin de les comparer, que ce soit en terme de qualité d’apprentissage, de qualité d’inférence, de temps de calcul, et de capacité de stockage / Reliability analysis is an integral part of system design and operation, especially for systems running critical applications. Recent works have shown the interest of using Bayesian Networks in the field of reliability, for modeling the degradation of a system. The Graphical Duration Models are a specific case of Bayesian Networks, which make it possible to overcome the Markovian property of dynamic Bayesian Networks. They adapt to systems whose sojourn-time in each state is not necessarily exponentially distributed, which is the case for most industrial applications. Previous works, however, have shown limitations in these models in terms of storage capacity and computing time, due to the discrete nature of the sojourn time variable. A solution might be to allow the sojourn time variable to be continuous. According to expert opinion, sojourn time variables follow a Weibull distribution in many systems. The goal of this thesis is to integrate sojour time variables following a Weibull distribution in a Graphical Duration Model by proposing a new approach. After a presentation of the Bayesian networks, and more particularly graphical duration models, and their limitations, this report focus on presenting the new model allowing the modeling of the degradation process. This new model is called Weibull Hybrid Graphical Duration Model. An original algorithm allowing inference in such a network has been deployed. Various so built databases allowed to learn on one hand a Graphical Duration Model, and on an other hand a Graphical Duration Model Hybrid - Weibull, in order to compare them, in term of learning quality, of inference quality, of compute time, and of storage space

[pt] Imagens de sensoriamento remoto permitem o monitoramento e mapeamento de culturas de maneira precisa, apoiando práticas de agriculturaeficientes e sustentáveis com o objetivo de garantir a segurança alimentar.No entanto, a identificação do tipo de cultura a partir de dados de sensoriamento remoto em regiões tropicais ainda são consideradas tarefas comalto grau de dificuldade. As favoráveis condições climáticas permitem o uso,planejamento e o manejo da terra com maior flexibilidade, o que implica emculturas com dinâmicas mais complexas. Além disso, a presença constantede nuvens dificulta o uso de imagens ópticas, tornando as imagens de radar uma alternativa interessante para o mapeamento de culturas em regiõestropicais. Os modelos de campos aleatórios condicionais (CRFs) têm sidousados satisfatoriamente para explorar o contexto temporal e espacial naclassificação de imagens de sensoriamento remoto. Estes modelos oferecemuma alta precisão na classificação, no entanto, dependem de atributos extraídos manualmente com base em conhecimento especializado do domínio.Neste contexto, os métodos de aprendizado profundo, tais como as redesneurais convolucionais (CNNs), provaram ser uma alternativa robusta paraa classificação de imagens de sensoriamento, pois podem aprender atributosótimos diretamente dos dados. Este trabalho apresenta um modelo híbridobaseado em aprendizado profundo e CRF para o reconhecimento de culturas em áreas de regiões tropicais caracterizadas por ter uma dinâmicaespaço–temporal complexa. O framework proposto consiste em dois módulos: uma CNNs que modela o contexto espacial e temporal dos dados deentrada, e o CRF que modela a dinâmica temporal considerando a dependência entre rótulos para datas adjacentes. Estas dependências podem seraprendidas ou desenhadas por um especialista nas práticas de agriculturalocal. Comparações entre diferentes variantes de como modelar as transiçõestemporais são apresentadas usando sequências de imagens SAR de duas municipalidades no Brasil. Os experimentos mostraram melhorias significativasatingindo ate 30 por cento no F1 score por classe e ate 12 por cento no F1 score medio em relação ao modelo de base que não inclui dependências temporais duranteo processo de aprendizagem. / [en] Remote sensing imagery enables accurate crop mapping and monitoring, supporting efficient and sustainable agricultural practices to ensure food security. However, accurate crop type identification and crop area estimation from remote sensing data in tropical regions are still challenging tasks. Compared to the characteristic conditions of temperate regions, the more favorable weather conditions in tropical regions permit higher flexibility in land use, planning, and management, which implies complex crop dynamics. Moreover, the frequent cloud cover prevents the use of optical data during large periods of the year, making SAR data an attractive alternative for crop mapping in tropical regions. To exploit both spatial and temporal contex, conditional random fields (CRFs) models have been used successfully in the classification of RS imagery. These approaches deliver high accuracies; however, they rely on features engineering manually designed based on domain-specific knowledge. In this context, deep learning methods such as convolutional neural networks (CNNs) proved to be a robust alternative for remote sensing image classification, as they can learn optimal features and classification parameters directly from raw data. This work introduces a novel end-to-end hybrid model based on deep learning and conditional random fields for crop recognition in areas characterized by complex spatio-temporal dynamics typical of tropical regions. The proposed framework consists of two modules: a CNN that models spatial and temporal contexts from the input data and a CRF that models temporal dynamics considering label dependencies between adjacent epochs. These dependencies can be learned or designed by an expert in local agricultural practices. Comparisons between data-driven and prior-knowledge temporal constraints are presented for two municipalities in Brazil, using multi-temporal SAR image sequences. The experiments showed significant improvements in per class F1 score of up to 30 percent and up to 12 percent in average F1 score against a baseline model that doesn t include temporal dependencies during the learning process.

No description available.

Nous proposons une approche pour détecter les sujets, les communautés d'intérêt non disjointes,l'expertise, les tendances et les activités dans des sites où le contenu est généré par les utilisateurs et enparticulier dans des forums de questions-réponses tels que StackOverFlow. Nous décrivons d'abordQASM (Questions & Réponses dans des médias sociaux), un système basé sur l'analyse de réseauxsociaux pour gérer les deux principales ressources d’un site de questions-réponses: les utilisateurs et lecontenu. Nous présentons également le vocabulaire QASM utilisé pour formaliser à la fois le niveaud'intérêt et l'expertise des utilisateurs. Nous proposons ensuite une approche efficace pour détecter lescommunautés d'intérêts. Elle repose sur une autre méthode pour enrichir les questions avec un tag plusgénéral en cas de besoin. Nous comparons trois méthodes de détection sur un jeu de données extrait dusite populaire StackOverflow. Notre méthode basée sur le se révèle être beaucoup plus simple et plusrapide, tout en préservant la qualité de la détection. Nous proposons en complément une méthode pourgénérer automatiquement un label pour un sujet détecté en analysant le sens et les liens de ses mots-clefs.Nous menons alors une étude pour comparer différents algorithmes pour générer ce label. Enfin, nousétendons notre modèle de graphes probabilistes pour modéliser conjointement les sujets, l'expertise, lesactivités et les tendances. Nous le validons sur des données du monde réel pour confirmer l'efficacité denotre modèle intégrant les comportements des utilisateurs et la dynamique des sujets / We propose an approach to detect topics, overlapping communities of interest, expertise, trends andactivities in user-generated content sites and in particular in question-answering forums such asStackOverFlow. We first describe QASM (Question & Answer Social Media), a system based on socialnetwork analysis to manage the two main resources in question-answering sites: users and contents. Wealso introduce the QASM vocabulary used to formalize both the level of interest and the expertise ofusers on topics. We then propose an efficient approach to detect communities of interest. It relies onanother method to enrich questions with a more general tag when needed. We compared threedetection methods on a dataset extracted from the popular Q&A site StackOverflow. Our method basedon topic modeling and user membership assignment is shown to be much simpler and faster whilepreserving the quality of the detection. We then propose an additional method to automatically generatea label for a detected topic by analyzing the meaning and links of its bag of words. We conduct a userstudy to compare different algorithms to choose the label. Finally we extend our probabilistic graphicalmodel to jointly model topics, expertise, activities and trends. We performed experiments with realworlddata to confirm the effectiveness of our joint model, studying the users’ behaviors and topicsdynamics