Les problématiques abordées dans ma thèse sont de définir une adaptation unifiée entre la sélection des documents et les stratégies de recherche de la réponse à partir du type des documents et de celui des questions, intégrer la solution au système de Questions-Réponses (QR) RITEL du LIMSI et évaluer son apport. Nous développons et étudions une méthode basée sur une approche de Recherche d’Information pour la sélection de documents en QR. Celle-ci s’appuie sur un modèle de langue et un modèle de classification binaire de texte en catégorie pertinent ou non pertinent d’un point de vue QR. Cette méthode permet de filtrer les documents sélectionnés pour l’extraction de réponses par un système QR. Nous présentons la méthode et ses modèles, et la testons dans le cadre QR à l’aide de RITEL. L’évaluation est faite en français en contexte web sur un corpus de 500 000 pages web et de questions factuelles fournis par le programme Quaero. Celle-ci est menée soit sur des documents complets, soit sur des segments de documents. L’hypothèse suivie est que le contenu informationnel des segments est plus cohérent et facilite l’extraction de réponses. Dans le premier cas, les gains obtenus sont faibles comparés aux résultats de référence (sans filtrage). Dans le second cas, les gains sont plus élevés et confortent l’hypothèse, sans pour autant être significatifs. Une étude approfondie des liens existant entre les performances de RITEL et les paramètres de filtrage complète ces évaluations. Le système de segmentation créé pour travailler sur des segments est détaillé et évalué. Son évaluation nous sert à mesurer l’impact de la variabilité naturelle des pages web (en taille et en contenu) sur la tâche QR, en lien avec l’hypothèse précédente. En général, les résultats expérimentaux obtenus suggèrent que notre méthode aide un système QR dans sa tâche. Cependant, de nouvelles évaluations sont à mener pour rendre ces résultats significatifs, et notamment en utilisant des corpus de questions plus importants. / This thesis aims at defining a unified adaptation of the document selection and answer extraction strategies, based on the document and question types, in a Question-Answering (QA) context. The solution is integrated in RITEL (a LIMSI QA system) to assess the contribution. We develop and investigate a method based on an Information Retrieval approach for the selection of relevant documents in QA. The method is based on a language model and a binary model of textual classification in relevant or irrelevant category. It is used to filter unusable documents for answer extraction by matching lists of a priori relevant documents to the question type automatically. First, we present the method along with its underlying models and we evaluate it on the QA task with RITEL in French. The evaluation is done on a corpus of 500,000 unsegmented web pages with factoid questions provided by the Quaero program (i.e. evaluation at the document level or D-level). Then, we evaluate the methodon segmented web pages (i.e. evaluation at the segment level or S-level). The idea is that information content is more consistent with segments, which facilitates answer extraction. D-filtering brings a small improvement over the baseline (no filtering). S-filtering outperforms both the baseline and D-filtering but not significantly. Finally, we study at the S-level the links between RITEL’s performances and the key parameters of the method. In order to apply the method on segments, we created a system of web page segmentation. We present and evaluate it on the QA task with the same corpora used to evaluate our document selection method. This evaluation follows the former hypothesis and measures the impact of natural web page variability (in terms of size and content) on RITEL in its task. In general, the experimental results we obtained suggest that our IR-based method helps a QA system in its task, however further investigations should be conducted – especially with larger corpora of questions – to make them significant.

O presente estudo analisa a rede de fóruns modulares produzidos pelo extinto Departamento de Obras Públicas do Estado de São Paulo - DOP - entre os anos de 1969 e 1975. Tem como objetivo avaliar se os projetos padronizados (ou normatizados), F, F1, F2 e F3, respondiam aos parâmetros convencionais de produção do Departamento, ou se existiu uma aproximação programática com os ideários da arquitetura moderna, envolvendo, quer do ponto de vista da sua dimensão social, quer da sua componente estética, a racionalização da produção e/ou a industrialização da construção, solicitando a modulação de todos seus elementos e componentes - constituindo-se, assim, como possibilidade da ampliação quantitativa de um bem, tal como a habitação, ou de uma rede de serviços: educação, saúde e, no caso específico, o acesso ao judiciário. Nesse sentido, a análise pondera sobre a continuidade da produção de fóruns de justiça, a partir do Plano de Ação Carvalho Pinto - PAGE, (1959-1963), período no qual os edifícios públicos foram projetados por arquitetos, em boa parte, comprometidos com a arquitetura moderna em São Paulo. O mesmo período tornou-se significativo para a produção moderna no Estado e, ao mesmo tempo, renovou-se a produção dos edifícios públicos do judiciário. Por último, nota-se que o DOP teve sua estrutura operacional e administrativa reformulada sob a orientação da equipe técnica do Governo do Estado. / The present study analyzes the network of modular forums produced by the former Department of Public Works of the São Paulos State - DOP -between the years 1969 and 1975. Assessing the extent to which these standardized designs (or normalized) \'F, F1, F1, F2 e F3\' responded to conventional parameters of the production department, in implementation of a network of services: education, health, and, in this case, access to justice, with some programmatic approach with the ideals of modern architecture, involving the social dimension , the aesthetic component, the standardization of projects, not considering the rationalization of production and/or the industrialization of construction. Thus, the analysis sticks to the continuities and ruptures of the production of Forums from Plano de Ação Carvalho Pinto - PAGE -(1959/1963), time in which public buildings were designed by architects largely committed to modern architecture in São Paulo, and that constituted a significant moment of modern production in the state, while the building of public buildings of the judiciary section was renewed and the DOP had its operational and management structure reformulated under the guidance of the technical staff of the State Government.

O presente estudo analisa a rede de fóruns modulares produzidos pelo extinto Departamento de Obras Públicas do Estado de São Paulo - DOP - entre os anos de 1969 e 1975. Tem como objetivo avaliar se os projetos padronizados (ou normatizados), F, F1, F2 e F3, respondiam aos parâmetros convencionais de produção do Departamento, ou se existiu uma aproximação programática com os ideários da arquitetura moderna, envolvendo, quer do ponto de vista da sua dimensão social, quer da sua componente estética, a racionalização da produção e/ou a industrialização da construção, solicitando a modulação de todos seus elementos e componentes - constituindo-se, assim, como possibilidade da ampliação quantitativa de um bem, tal como a habitação, ou de uma rede de serviços: educação, saúde e, no caso específico, o acesso ao judiciário. Nesse sentido, a análise pondera sobre a continuidade da produção de fóruns de justiça, a partir do Plano de Ação Carvalho Pinto - PAGE, (1959-1963), período no qual os edifícios públicos foram projetados por arquitetos, em boa parte, comprometidos com a arquitetura moderna em São Paulo. O mesmo período tornou-se significativo para a produção moderna no Estado e, ao mesmo tempo, renovou-se a produção dos edifícios públicos do judiciário. Por último, nota-se que o DOP teve sua estrutura operacional e administrativa reformulada sob a orientação da equipe técnica do Governo do Estado. / The present study analyzes the network of modular forums produced by the former Department of Public Works of the São Paulos State - DOP -between the years 1969 and 1975. Assessing the extent to which these standardized designs (or normalized) \'F, F1, F1, F2 e F3\' responded to conventional parameters of the production department, in implementation of a network of services: education, health, and, in this case, access to justice, with some programmatic approach with the ideals of modern architecture, involving the social dimension , the aesthetic component, the standardization of projects, not considering the rationalization of production and/or the industrialization of construction. Thus, the analysis sticks to the continuities and ruptures of the production of Forums from Plano de Ação Carvalho Pinto - PAGE -(1959/1963), time in which public buildings were designed by architects largely committed to modern architecture in São Paulo, and that constituted a significant moment of modern production in the state, while the building of public buildings of the judiciary section was renewed and the DOP had its operational and management structure reformulated under the guidance of the technical staff of the State Government.

Des études ont récemment mis en évidence le rôle des lipopolysaccharides (LPS) des bactéries à Gram négatif de la flore intestinale dans le processus de l’inflammation conduisant à l’obésité et au diabète de type 2.Le présent travail est réalisé dans le cadre d’une collaboration entre les équipes du Dr. Caroff (U. Paris-Sud, Orsay) et du Pr. Zhao (U. Jiao Tong, Shanghai). Les expériences présentées ont été réalisées lors de séjours dans les deux laboratoires.Il a été démontré en Chine que des bactéries Sulfato-réductrices (SRB) à Gram négatif étaient présentes en plus forte proportion dans la flore intestinale chez les souris suivant un régime gras. Les mêmes résultats ont été observés chez l'homme. L’hypothèse selon laquelle des SRB seraient à l’origine de grandes quantités d’endotoxines chez les obèses et les patients diabétiques a été émise. Plusieurs souches de SRB isolées de la flore intestinale humaine d’un sujet sain et d’un sujet diabétique ont été cultivées en Chine. Des études de relation structure/activité des LPS isolés de ces bactéries ont été réalisées dans le laboratoire Français pour déterminer leur rôle dans le développement des maladies métaboliques. Les souches isolées des deux sujets ont pu être classées dans le genre Desulfovibrio. Les LPS correspondants ont été extraits et purifiés par des méthodes mises au point dans l’équipe d’Orsay. La structure chimique a été élucidée par les méthodes suivantes : Electrophorèse, Chromatographie sur couche mince, Chromatographie en phase gazeuse et Spectrométrie de masse MALDI. C’est ainsi que des spectres de masse ont été obtenus et que la structure des lipides A, principes actifs des LPS, isolés de SRB a été décrite pour la première fois. Les activités biologiques testées (TNFα, IL-6) varient en fonction du nombre d’acides gras présents. Les LPS de SRB du patient sain ont une structure variable (Smooth versus Rough) en fonction de la quantité de fer présent dans le milieu, et ceux isolés du patient diabétique présentent des structures atypiques qui ne sont pas toutes inflamogènes. Une molécule membranaire inconnue, que nous avons nommée « Glycosyl’X » était co-extraite avec les LPS. Elle joue apparemment un rôle important dans la croissance des SRB et a été étudiée après des étapes de purification complexes. Les structures et le pouvoir inflammatoire de ces molécules dont la structure varie avec les souches, et qui chélatent le fer, ont été étudiées. Elles sont de nature principalement osidique et fixées à la membrane. La proportion de ces molécules par rapport aux LPS varie avec la quantité de fer disponible dans le milieu. Un milieu riche en fer favorise la croissance des Desulfovibrio portant les Glycosyl’X qui n’ont pas de pouvoir inflammatoire eux-mêmes, mais entrent en compétition avec les LPS, modulant ainsi indirectement l’activité de ces derniers. L’augmentation du nombre de Desulfovibrio conduisant à l’augmentation des molécules Glycosyl’X pourrait aussi moduler positivement (par présentation) ou négativement (par élimination des bactéries) l’adsorption du fer dans les intestins dont l’équilibre est essentiel pour l’homéostasie métabolique.Par ailleurs, la croissance des Desulfovibrio augmente la production d’Hydrogène Sulfuré connu pour son action délétère sur les cellules. Nous favorisons l’hypothèse selon laquelle son action sur la disjonction des cellules épithéliales permettrait le passage des différents LPS relargués par la flore Gram-négative intestinale, et même des bactéries entières, vers la circulation sanguine. / Recent studies have highlighted the role of lipopolysaccharide (LPS) in the intestinal flora (gut microbiota) which could contribute to the inflammation process leading to obesity and type 2 diabetes. This thesis is part of a collaborative project between the laboratories of Dr. Caroff (U. Paris -Sud, Orsay, France) and Prof. Zhao (U. Jiao Tong , Shanghai, China). It has been shown by Pr.Zhao’s team in 2010 that the Sulfate -Reducing Bacteria (SRB) were presented in greater proportion in the intestinal mice flora following a fat diet compared to mice following a normal diet. The same results were observed in humans. The starting hypothesis was that SRB could produce a large amount of endotoxin in obese and diabetic patients and play a role in the development of metabolic diseases. Several SRB strains isolated from the human intestinal flora of a healthy subject and of a diabetic subject were grown in the Chinese laboratory. Studies of their LPS structure / activity relationships were carried out in the French laboratory. The aim of this study was to determine their roles in the development of metabolic diseases.Strains isolated from the two subjects could be classified in the Desulfovibrio genus. The corresponding LPS were extracted and purified by the methods developed in the French laboratory. The chemical structure was elucidated by the following methods: Electrophoresis, Thin layer chromatography, Gas chromatography and MALDI mass spectrometry. The mass spectra were obtained and the structure of lipid A, the active part of LPS isolated from SRB was described here for the first time. The biological activities test (TNFα, IL-6) vary depending on the number of fatty acids present in their lipid A structure. The LPS of SRB isolated from the healthy patient had a variable structure (Smooth versus Rough) depending on the amount of iron present in the medium, and those isolated from diabetic patients had atypical structures are not all inflamogenic .An unknown membrane molecule, which we named "Glycosyl'X" was co-extracted with the LPS. It apparently plays an important role in the growth of SRB was investigated after complex purification steps. The structures and the inflammatory power of these molecules variying with strains chelating iron were studied. They are mainly of glycosidic nature and linked to the bacterial membrane.The proportion of these molecules relatively to LPS varies with the amount of iron in the medium. An environment rich in iron promotes the growth of Desulfovibrio Glycosyl'X, molecules but competes with LPS and indirectly modulates the activity of the latter. The increase number of Desulfovibrio leading to increased Glycosyl'X molecules may also modulate positively (by presentation) or negatively (by killing bacteria) the absorption of iron in the intestines which balance is essential for metabolic homeostasis.Furthermore, the growth of Desulfovibrio increasing the production of Hydrogen Sulfide is known for its deleterious effects on the cells. We favor the hypothesis that its action on the separation of epithelial cells favors the passage of different LPS released by the Gram- negative of intestinal flora and even whole cell bacteria into the bloodstream.

Τρεις ευρέως χρησιμοποιούμενες μέθοδοι τυποποίησης, μια βιοχημική (API20NE), μια φαινοτυπική (SDS-PAGE) και μια μοριακή (RAPD) χρησιμοποιήθηκαν για την ταυτοποίηση και ταξινόμηση 160 περιβαλλοντικών ψευδομονάδων που απομονώθηκαν από το υδάτινο περιβάλλον της Νοτιοδυτικής Ελλάδας και συγκεκριμένα από εμφιαλωμένα νερά (46%), νερά δικτύου ύδρευσης (16%), κολυμβητικών δεξαμενών (9%) και θαλασσών (29%). Οι ψευδομονάδες ταυτοποιήθηκαν με βάση το βιοχημικό τους αποτύπωμα δια μέσου του συστήματος ΑΡΙ20ΝΕ, και στη συνέχεια υποβλήθηκαν σε ηλεκτροφόρηση των ολικών πρωτεϊνών τους (SDS-PAGE) και σε ανάλυση του γενετικού τους υλικού με τη μέθοδο RAPD (Random Amplified Polymorphic DNAs) με χρήση δύο διαφορετικών δεκαμερών εκκινητών (primers). Το σύστημα API20NE ταυτοποίησε το 88% των στελεχών διακρίνοντας 14 ομάδες-είδη, ενώ η SDS-PAGE ταξινόμησε το 98.1% σε 20 ομάδες και η RAPD το 94% των στελεχών σε 22 και 34 ομάδες, με εκκινητή τον OPA-13 και τον OPD-13 αντίστοιχα. Η ταξινόμηση προέκυψε με εφαρμογή της ανάλυσης κατά συστάδες (cluster analysis) των αποτυπωμάτων (πρωτεϊνικών και γενετικών) που παρήγαγαν τα στελέχη. Τα 20 στελέχη που δεν ταυτοποιήθηκαν σε επίπεδο είδους με το API20NE, ταξινομήθηκαν με την SDS-PAGE σε ποσοστό 100%, ενώ με την RAPD σε ποσοστό 90%. Οι τρεις μέθοδοι συγκρίθηκαν ως προς την επαναληψιμότητα (reproducibility), την ικανότητα τυποποίησης (typeability) και τη διακριτική ικανότητα (discriminatory power). Την μεγαλύτερη επαναληψιμότητα έδωσαν το API20NE και η RAPD με τον εκκινητή OPA-13, την μεγαλύτερη ικανότητα τυποποίησης η SDS-PAGE, ενώ τη μεγαλύτερη διακριτική ικανότητα έδωσε η RAPD με τον εκκινητή OPD-13. Η πλέον σωστή ταξινόμηση, όπως προέκυψε από τη διακριτή ανάλυση, επιτεύχθη με τη μέθοδο SDS-PAGE. Η παρούσα εργασία αποδεικνύει ότι τα βιοχημικά συστήματα ταυτοποίησης (όπως το API20NE) μπορούν να χρησιμοποιηθούν με αξιοπιστία μόνο για αδρή αναγνώριση των περιβαλλοντικών ψευδομονάδων. Πληροφορίες σε βάθος για την ταυτότητα και τη φύση τους μπορούν να εξαχθούν με τη περαιτέρω εφαρμογή ηλεκτροφορητικών και μοριακών μεθόδων. Δεδομένης της ευρείας διασποράς, της ετερογένειας και της, έστω και δυνητικής, παθογόνου δράσης των ψευδομονάδων, είναι σημαντικό, από πλευράς δημόσιας υγείας, ο προσδιορισμός της ταυτότητάς τους να γίνεται με συνδυασμένη εφαρμογή βιοχημικών, ηλεκτροφορητικών και μοριακών μεθόδων ώστε να καθίσταται δυνατή η αναγνώριση στελεχών που μπορούν να αποτελέσουν αιτιολογικούς παράγοντες ασθενειών, ιδιαίτερα σε ομάδες υψηλού κινδύνου. / Three broadly used typing techniques, one biochemical (API20NE), one phenotypic (SDS-PAGE) and one molecular (RAPD), were employed for the identification and taxonomy of 160 environmental pseudomonas isolated from the aquatic environment in Southwestern Greece. In particular, the isolates were obtained from bottled waters (46%), potable waters (16%), waters from swimming pools (9%) and seawaters (29%). The isolates were identified by the system API20NE and then subjected to whole-cell protein electrophoresis (SDS-PAGE) and Random Amplified Polymorphic DNAs (RAPD) using two 10-mer primers. The API20NE system identified 88% of the whole bacterial population and classified them in 14 species, while SDS-PAGE classified 98.1% of the isolates in 20 groups and RAPD classified 94% of the strains in 22 groups using the primer OPA-13 and 34 groups using the primer OPD-13. The classification was achieved by applying cluster analysis in the protein or RAPD fingerprints of the isolates. Twenty isolates that could not be identified by the API20NE system, at least to the species level, were classified by the SDS-PAGE and the RAPD in a percentage of 100% and 90%, respectively. The reproducibility, typeability and discriminatory power of the three methods were compared to evaluate their application. The API20NE and the RAPD assay with primer OPA-13 showed better reproducibility in comparison with the other methods; the higher typeability was achieved by the SDS-PAGE assay while the higher discriminatory power was that obtained by the RAPD method with the primer OPD-13. The SDS-PAGE gave the higher percentage of “correctly classified” isolates, as it was assessed by discriminant analysis. This study shows that the rapid identification systems, such as the API20NE, may be reliable only for a rough characterization of environmental Pseudomonas. In order to acquire further information about their identities, other phenotypic and molecular techniques have to be applied. Given the ubiquity, heterogeneity and pathogenicity, either established or potential, of the environmental pseudomonas it is important, from a public health point of view, to monitor the identities of environmental Pseudomonas isolates using the combination of specific methods, so as to be possible for strains, which can serve as causative agents of diseases, especially in high risk population, to be recognizable.

No description available.

No description available.

The two related putative cytochrome c oxidase (COX) assembly factors HCC1 and HCC2 from Arabidopsis thaliana are Homologs of the yeast Copper Chaperones Sco1p and Sco2p. The hcc1 null mutation was previously shown to be embryo lethal while the disruption of the HCC2 gene function had no obvious effect on plant development, but increased the expression of stress-responsive genes. Both HCC1 and HCC2 contain a thioredoxin domain, but only HCC1 carries a Cu-binding motif also found in Sco1p and Sco2p. In order to investigate the physiological implications suggested by this difference, various hcc1 and hcc2 mutants were generated and analyzed. The lethality of the hcc1 knockout mutation was rescued by complementation with the HCC1 gene under the control of the embryo-specific promoter ABSCISIC ACID INSENSITIVE 3. However, the complemented seedlings did not grow into mature plants, underscoring the general importance of HCC1 for plant growth. The HCC2 homolog was shown to localize to mitochondria like HCC1, yet the function of HCC2 is evidently different, because two hcc2 knockout lines developed normally and exhibited only mild growth suppression compared with the wild type (WT). However, hcc2 knockouts were more sensitive to UV-B treatment than the WT. Complementation of the hcc2 knockout with HCC2 rescued the UV-B-sensitive phenotype. In agreement with this, exposure of wild-type plants to UV-B led to an increase of HCC2 transcripts. In order to corroborate a function of HCC1 and HCC2 in COX biogenesis, COX activity of hcc1 and hcc2 mutants was compared. While the loss of HCC2 function had no significant effect on COX activity, the disruption of one HCC1 gene copy was enough to suppress respiration by more than half compared with the WT. Therefore, we conclude that HCC1 is essential for COX function, most likely by delivering Cu to the catalytic center. HCC2, on the other hand, seems to be involved directly or indirectly in UV-B-stress responses.

This diploma thesis deals with developing generator for single-page applications. Before developing the application it was necessary to identify problem areas that prevent the development and describe tools that make it easy to create, test, maintain, and deploy single-page applications. Based on the obtained information, the generator Create Sbspa is designed and implemented to efficiently create single-page applications and help to eliminate development problems. It generates configuration and code from semantic templates. The generator is available through a user interface that splits the templates into the groups by applicability. The generator was designed with the need for simplicity and clarity to enable efficient integration with new features. This work also includes design and implementation of the example app which shows features and benefits of the generator.

This thesis aims at creation of new datasets for text recognition machine learning tasks and experiments with convolutional neural networks on these datasets. It describes architecture of convolutional nets, difficulties of recognizing text from photographs and contemporary works using these networks. Next, creation of annotation, using Tesseract OCR, for dataset comprised from photos of document pages, taken by mobile phones, named Mobile Page Photos. From this dataset two additional are created by cropping characters out of its photos formatted as Street View House Numbers dataset. Dataset Mobile Nice Page Photos Characters contains readable characters and Mobile Page Photos Characters adds hardly readable and unreadable ones. Three models of convolutional nets are created and used for text recognition experiments on these datasets, which are also used for estimation of annotation error.