Cette thèse traite de la conception de méthodes automatisées ou semi-automatisées pour détecter et identifier des algorithmes cryptographiques dans des programmes compilés en langage machine. La première méthode proposée a pour but l'identification de primitives symétriques. L'implémentation en langage machine d'une primitive symétrique, assimilée à une suite d'instructions, est représentée par un graphe. Sous cette forme, le code est modifié à l'aide de règles de réécriture tout en préservant une certaine notion de sémantique lors d'une phase dite de normalisation. L'objectif est de faire émerger des expressions communes à différentes implémentations d'une même primitive. Ces expressions servent alors de base à la création de signatures efficaces. La recherche de ces signatures s'effectue à l'aide d'un algorithme énumérant les isomorphismes de sous-graphe. La seconde méthode, conçue en complément de la première, produit une représentation synthétique facilitant l'identification des modes opératoires. Cette représentation se définit comme le plus petit sous-graphe préservant les distances entre des sous-ensembles de nœuds précédemment identifiés comme étant les paramètres d'entrée et de sortie des primitives impliquées. / This thesis is about the design of automatic or semi-automatic methods to detect and identify cryptographic algorithms inside programs compiled into machine code. Two methods are presented. The objective of the first method is to identify cryptographic primitives. A machine code implementation of a cryptographic primitive, regarded as a sequence of instructions, is represented by a graph structure. During a normalization phase, a set of rewrite rules is used to modify this graph representation while preserving a specific notion of semantics. The goal is to converge towards expressions which are shared across several implementations of the same primitive. We use these expressions as a basis to create efficient signatures. A subgraph isomorphism enumeration algorithm is used to search for signatures. The second method is built on top of the first one. It produces a synthetic representation designed to help in the identification of modes of operation. This synthetic representation is defined as the smallest subgraph which preserve distances between sets of vertices previously identified as the input and output parameters of the primitives involved within the mode of operation.

The thesis aims on methods and strategies usable for information retrieval on the internet as we know it today and points at how the requirements on knowledge of end-user and abilities to use different information sources have been changing; in other words it points the increasing demands on information literacy. The document is divided into three parts. The first one briefly provides an introduction to strategies of information retrieval. The second part deals on theoretic basis with tools and information sources for searching on the both surface and deep web. The author also focuses on social web and its special attributes from the searcher's point of view. The third part represents a practical utilization of knowledge from previous parts. Practical part demonstrates on model examples a nature and kind of information we can retrieve from particular resources.

No description available.

Esta pesquisa apresenta o desenvolvimento de método de busca e recuperação de informações em bases de dados especializadas para produção do conhecimento científico na área da Saúde, com ênfase na Saúde Baseada em Evidências. Recorremos, neste trabalho, a diferentes metodologias considerando as especificidades de cada etapa: pesquisa exploratória, método hipotético dedutivo e estudo de caso empírico qualitativo. Mobilizamos os fundamentos teórico-metodológicos da Ciência da Informação e da Saúde nos domínios da Organização e Recuperação da Informação e do Conhecimento, Web Semântica, Saúde Baseada em Evidências e Metodologia Científica, assim como realizamos dois experimentos: estudo de caso em Ortopedia e Traumatologia no sentido de identificar e estabelecer critérios para busca, recuperação, organização e seleção de informações de modo que possam integrar parte da metodologia de trabalhos científicos na área da Saúde; e análise dos tipos de busca e recuperação e dos mapeamentos entre Sistemas de Organização do Conhecimento (SOC) propostos no Metatesauro no escopo da Unified Medical Language System (UMLS) da US National Library of Medicine e no BioPortal da National Center for Biomedical Ontology, ambos na área biomédica. O UMLS disponibiliza acesso a 151 SOC, e o BioPortal, um conjunto de 302 ontologias. Apresentam-se propostas para construção de estratégias de busca com uso de Sistemas de Organização do Conhecimento mapeados e interoperados, bem como para realização de pesquisas bibliográficas para elaboração de trabalhos científicos na área da Saúde. / This research presents the development of method for search and information retrieval in specialized databases aiming the production of scientific knowledge in healthcare, with emphasis on Evidence-Based Health. We have used, in this work, different techniques considering the specificities of each stage: exploratory research, hypothetical deductive method and qualitative empirical case study. It mobilizes the theoretical and methodological foundations in Information Science and Health, appling them to areas as knowledge organization and information retrieval, Semantic Web, Evidence-Based Health and Scientific Methodology. Two experiments were performed: a case study in Orthopedics and Traumatology in order to identify and establish criterions for search, retrieval, organization and selection of information, so that these criterions can integrate part of the methodology of scientific work in healthcare; and analysis of kinds of search and retrieval and mappings on Knowledge Organization Systems-KOS available in Metathesaurus, considering the scope of the Unified Medical Language System (UMLS) of the US National Library of Medicine (NLM), and in the BioPortal National Center for Biomedical Ontology, both in the biomedical field. The UMLS provides access to 151 KOS, and the BioPortal provides a set of 302 ontologies. We presented proposals for construction of search strategies by using Knowledge Organization System mapped and interoperate as well as for conducting literature searches for preparation of scientific papers in healthcare.

Esta pesquisa apresenta o desenvolvimento de método de busca e recuperação de informações em bases de dados especializadas para produção do conhecimento científico na área da Saúde, com ênfase na Saúde Baseada em Evidências. Recorremos, neste trabalho, a diferentes metodologias considerando as especificidades de cada etapa: pesquisa exploratória, método hipotético dedutivo e estudo de caso empírico qualitativo. Mobilizamos os fundamentos teórico-metodológicos da Ciência da Informação e da Saúde nos domínios da Organização e Recuperação da Informação e do Conhecimento, Web Semântica, Saúde Baseada em Evidências e Metodologia Científica, assim como realizamos dois experimentos: estudo de caso em Ortopedia e Traumatologia no sentido de identificar e estabelecer critérios para busca, recuperação, organização e seleção de informações de modo que possam integrar parte da metodologia de trabalhos científicos na área da Saúde; e análise dos tipos de busca e recuperação e dos mapeamentos entre Sistemas de Organização do Conhecimento (SOC) propostos no Metatesauro no escopo da Unified Medical Language System (UMLS) da US National Library of Medicine e no BioPortal da National Center for Biomedical Ontology, ambos na área biomédica. O UMLS disponibiliza acesso a 151 SOC, e o BioPortal, um conjunto de 302 ontologias. Apresentam-se propostas para construção de estratégias de busca com uso de Sistemas de Organização do Conhecimento mapeados e interoperados, bem como para realização de pesquisas bibliográficas para elaboração de trabalhos científicos na área da Saúde. / This research presents the development of method for search and information retrieval in specialized databases aiming the production of scientific knowledge in healthcare, with emphasis on Evidence-Based Health. We have used, in this work, different techniques considering the specificities of each stage: exploratory research, hypothetical deductive method and qualitative empirical case study. It mobilizes the theoretical and methodological foundations in Information Science and Health, appling them to areas as knowledge organization and information retrieval, Semantic Web, Evidence-Based Health and Scientific Methodology. Two experiments were performed: a case study in Orthopedics and Traumatology in order to identify and establish criterions for search, retrieval, organization and selection of information, so that these criterions can integrate part of the methodology of scientific work in healthcare; and analysis of kinds of search and retrieval and mappings on Knowledge Organization Systems-KOS available in Metathesaurus, considering the scope of the Unified Medical Language System (UMLS) of the US National Library of Medicine (NLM), and in the BioPortal National Center for Biomedical Ontology, both in the biomedical field. The UMLS provides access to 151 KOS, and the BioPortal provides a set of 302 ontologies. We presented proposals for construction of search strategies by using Knowledge Organization System mapped and interoperate as well as for conducting literature searches for preparation of scientific papers in healthcare.