Search

We analyzed parameters influencing HIV-1 infectibility of cells of the monocyte/macrophage lineage (MO/MAC) isolated from different healthy donors. The proportion of in vitro-infected cells and replication kinetics in different donor MAC ranged from 0.03 to 99% p24 antigen-positive MAC and from undetectable RT activity up to 5 × 106 cpm/ml/90 min, respectively. As a quantitative measurement for HIV-1 susceptibility of donor MO/MAC, we determined TCID50 values of defined virus stocks which varied up to 3000-fold depending on the donor MAC used for titration. As host factors which may influence the viral infection we determined the expression of virus receptors CD4, CCR5, CXCR4, and CCR3 as well as the secretion of the natural ligands of CCR5, which altogether showed no correlation with HIV-1 infectibility of the cells. Moreover, other MO-derived secretory factors which might affect viral infection of these cells could be excluded. Furthermore, expression of maturation-related antigens CD14, CD16, HLA-DR, and MAX.1/CPM was determined. Analysis of the reverse transcription process revealed that restricted HIV-1 infection was reflected by highly reduced or even undetectable full-length HIV-1 DNA formation, although early and intermediate transcripts appeared, suggesting that viral replication is blocked after entry at the level of early reverse transcription.

MBE (Mail by Example) is a visual interface that provides advanced facilities for handling large volumes of electronic messages. It enables users to define ad hoc queries for retrieving messages, folders, or information about those. MBE is based on a “by-example” query style (QBE), to suit the requirements of typical users of email environments. The first evaluation of MBE revealed a generalized satisfaction towards its features.

This works presents an experience on the use of KDD (Knowledge Discovery in Databases) to identify and understandwhether curriculum revisions can affect students in a Brazilian university. Presently, there is no framework to define thenotion of impact caused by curriculum revisions, and the use of KDD can bring significant contributions, given the amountof data involved. The paper describes the analysis framework defined so far for measuring the impact of curriculumrevisions, and reports the results obtained after the analysis of students records related to five distinct degrees. The resultsobtained so far indicate that individual revisions quite often do not affect students, being sometimes even beneficial to them.However, considering the set of revisions they face during their academic lifetime, it is possible to generalize that manystudents are lightly harmed. This harm influences the number of extra-classes they have to take to fulfill the requirementsfor obtaining a given degree, but the time required to graduate is not affected by revisions.Keywords: KDD Application, Curriculum Revision.

Reusability is considered to be the key for achieving productivity and quality in software, and much has been claimed about the particular contributions of the object-oriented paradigm towards the achievement of these goals. Object-oriented frameworks are coarse-grained reuse units, composed of a set of classes specifically designed to be refined and used as a group. In this paper, we discuss the nature of frameworks necessary to build a particular type of systems, namely Decision Support Systems (DSS), and their organization in a generic OO DSS multi-layer architecture. DSS are systems intended to improve the effectiveness of decision making, but information technologies can only have a major impact on decision making if techniques allowing the easy and rapid development of DSS are available. Much benefit is expected in terms of easiness and rapidity of development by constructing DSS from domain-oriented reusable components, as well as in terms of quality of DSS in this way developed.

The qualitative relationship between preprotein translocases in the mitochondrial outer and inner membranes was determined by both a functional analysis and a determination of characteristic components of the translocases. Translocation contact sites of isolated mitochondria were saturated with intermediates of a matrix-targeted precursor of the β-subunit of the F1-ATPase (pF1β), and import of preproteins into the different mitochondrial subcompartments was monitored. A strong inhibition (75–95%) was observed for preproteins with an N-terminal matrix targeting signal, indicating that a significant portion of the contact sites was blocked by accumulated F1β. Insertion of preproteins into the outer membrane and import into the intermembrane space of preproteins without matrix targeting signals was inhibited by about 45%, indicating that functional outer membrane translocases were available despite saturation of contact sites. Similarly, import of members of the mitochondrial carrier family into the inner membrane was only partly inhibited (40–50%), demonstrating that functional Tim22 translocases were available to cooperate with the Tom machinery in the import of carrier proteins. The stoichiometry of Tom40, Tim23, and Tim22 in mitochondria was determined to be 5:1:0.22. We conclude that translocases of the outer membrane are present in excess over translocases of the inner membrane.

Interaction of the widely expressed Fas with its membrane-bound ligand (FasL) leads to rapid cell death via apoptosis. To avoid pathological tissue damage, the activity of FasL requires tight regulation. Here, we report that the Src homology 3 (SH3) domain of Fyn binds to the proline-rich cytoplasmic region of FasL. Binding of the SH3 domain occurs between amino acid residues 44–71 which contains several potential SH3 interaction sites. This binding is specific, as SH3 domains of Lck, Grb2 and ras-GAP bind only weakly or not at all. We suggest that FasL activity may be modulated by SH3 domains of the src-like Fyn kinase.

The aim of this study is to quantify the burden of chronic obstructive pulmonary disease (COPD) – incidence, prevalence, and mortality – and identify trends in Australia, Canada, France, Germany, Italy, Japan, The Netherlands, Spain, Sweden, the United Kingdom, and the United States of America. A structured literature search was performed (January 2000 to September 2010) of PubMed and EMBASE, identifying English-language articles reporting COPD prevalence, incidence, or mortality. Of 2838 articles identified, 299 full-text articles were reviewed, and data were extracted from 133 publications. Prevalence data were extracted from 80 articles, incidence data from 15 articles, and mortality data from 58 articles. Prevalence ranged from 0.2%–37%, but varied widely across countries and populations, and by COPD diagnosis and classification methods. Prevalence and incidence were greatest in men and those aged 75 years and older. Mortality ranged from 3–111 deaths per 100,000 population. Mortality increased in the last 30–40 years; more recently, mortality decreased in men in several countries, while increasing or stabilizing in women. Although COPD mortality increased over time, rates declined more recently, likely indicating improvements in COPD management. In many countries, COPD mortality has increased in women but decreased in men. This may be explained by differences in smoking patterns and a greater vulnerability in women to the adverse effects of smoking., Video abstract Point your SmartPhone at the code above. If you have a QR code reader the video abstract will appear. Or use: http://dvpr.es/Nl3LKT

Mitochondria are double-membrane bounded, energy-converting organelles within eukaryotic cells. They synthesize ATP by oxidative phosphorylation. Important progress in understanding their function was made when an isolation procedure for intact mitochondria (Hogeboom, Schneider & Palade, 1948) was developed. Time-lapse microcinematography (Johnson, Walsh & Chen, 1980; Tzagoloft, 1982) revealed that these organelles are remarkably mobile and able to constantly change their shape and are continuously dividing and fusing. Their motility is apparently due to an association with the cytoskeleton which may determine the unique orientation and distribution of mitochondria in various cell types. Mitochondria are not synthesized de novo; they arise by growth and division of pre-existing organelles (Luck, 1963, 1965). Mitochondria contain DNA (Schatz, Haslbrunner & Tuppy, 1964) and a complete system to carry out DNA replication, transcription, and protein synthesis (McLean et al., 1958; Tzagoloff, 1982); however, only a small number of proteins are encoded by mitochondrial DNA, namely a few subunits of the respiratory chain and the ATP synthase, and depending on the organism, one or a few proteins required for the expression of mitochondrial genes. All other mitochondrial proteins are encoded by nuclear genes, and thus have to be imported into the

Eukaryotic cells are functionally and morphologically subdivided into a number of membrane-bounded compartments. Each of the subcompartments, or organelles, has a specific protein composition according to its function in the cell. Mitochondria and, in plant cells, chloroplasts are unique in that they have their own genetic systems including DNA, ribosomes, tRNAs. Both organelles are surrounded by a double membrane and have extended internal membrane systems. According to the endosymbiotic hypothesis of the origin of mitochondria and chloroplasts both organelles are derived from prokaryotic endosymbionts in the ancestor of the eukaryotic cell (Margulis, 1970; Schwartz and Dayhoff, 1978). A large number of the original prokaryotic features are retained in the organelles. In the course of evolution, both organelles have lost most of their genetic information, which has been transferred to the nucleus (Schwartz and Dayhoff, 1978).

Neurotensin is a tridecapeptide with changing receptor expression in central and peripheral neural cells during ontogeny suggesting its potential use as a differentiation and tumor marker in neuroectodermal malignancies. We investigated the neurotensin levels in plasma samples of 58 patients with neuroblastoma using a sensitive radioimmunoassay. Elevated levels were found only in one stage III and in one stage IVs patient, while the neurotensin concentrations of 56 patients were in the range of control children. We conclude that plasma neurotensin reflects neither the differentiation nor the tumor status in children with neuroblastoma.

Hemophilia A is an X-chromosome linked bleeding disorder resulting from the absence or nonfunctional expression of coagulation factor VIII (FVIII). About 30% of severe hemophiliacs develop neutralizing antibodies (inhibitors) to FVIII upon treatment with exogenous factor preparations. Specificity of these antibodies is often restricted to functional determinants in the A2 and C2 domain. Phage displayed random peptide libraries were screened with various plasma samples of high titer inhibitor patients and inhibitor specific peptide ligands were selected. In silco mapping of consensus amino acid motifs among peptides revealed conformational epitopes in A2 and C2. Selected ligands partially restored FVIII activity. Equimolar combination of these ligands enhanced blocking of inhibitors in autologous and heterologous patients’ plasma. Peptide ligands were fused to the C-terminal multimerization domain of the C4bp alpha-chain and expressed as multimers in 293T cells. Peptide multimers revealed improved binding of anti-FVIII IgG and prolonged in vitro half-life in comparison to single synthetic peptides. Selected peptide ligands were combined in heteromultimers by co-transfection of respective vectors, resulting in molecules binding both A2- and C2- specific IgG and blocking up to 100% of antibody binding to FVIII. Those molecules could provide a basis for the generation of novel peptide-based therapeutic approaches.

PHOTOGRAPHY'S POSITION IN WESTERN CULTURE is embedded in an ideology of representation that regards it as simultaneously copying and constructing the world that it pictures. Photographic practice, which includes not only the taking of pictures, but also how photographs are looked at, thought about, saved, used, and re-used, illustrates the ways we resolve this apparent paradox. State institutions that rely on keeping records produce photographs that can be used as documents-as replicas of the world-with all apparent disinterest in how their pictorial records construct the objects and arena of the institutional gaze. Yet a closer examination of the routines an institution employs in gathering, selecting, and preserving its records reveals strategies of representation and conservation that create and perpetuate the world within the institutional boundaries, with implications reaching beyond its specific time and place. The photographs thus are joined in a mutually authenticating discourse that validates the institution's picture of its sphere at the same time that the photographs' ideological power is increased by their link to the institutional practices that have produced them. Museums established as official repositories of culture and history offer clear examples of this pattern. Charged with documenting and preserving that which is considered valuable, the museum has also become the institutionalized arbiter of value, determining what is worth collecting and saving for the future. Because photographing and archiving are primary ways of assigning value, they often occupy a central position in the museum's task of constructing and perpetuating a shared conception of a collective past. Since its founding in 1873, the Nordic Museum has been a central institution in the collection and documentation of Sweden's cultural history. Included in the museum's vast and intersecting collections of cultural artifacts are the pic

The boundary condition Re(Pw) = ϕ of an elliptic system in a bounded domain of the complex plane can be transformed into the Dirichlet boundary condition Re ω = ϕ (by means of the reversible substitution ω= Pw) if it is possible to extend the matrix-valued function P continuously to the whole of [Gbar] without violating the Lopatinski condition det P ≠ 0. Due to W. L. Wendland [8] this problem was solved by W. Tutschke [7] under special assumptions. Also in connection with elliptic systems W. L. Wendland in [8] used a homotopy from P to the unit matrix, constructed by piecewise Gauss elimination, but this method can fail as was shown by a counter-example. A complete but rather deep-lying result is contained in Bott's periodicity theorem, stating that with the functions P a - valued functional Grad P can be associated and the above mentioned problems are solvable if one has Grad P = 0. It is then easy to prove Bojarski's theorem [1], that Grad P is the number of revolutions of the vector det P during one c...

Photographs can offer insights into the structure and content of human communication, for they illustrate how people frame reality when translating it into a static mode for particular purposes. To continue to ignore photographs as resources for communication research in favor of modes of communication which appear more amenable to our understanding and which have established methods of analysis is to ignore the im pact of what we see on how we think.

Summary: Pseudomonas aeruginosa strain PAO1 was compared to PAO1 strains containing an IncP-1 R plasmid (RP1, R68, or R68.45) in an experimental mouse burn infection model. All R plasmids tested caused a 10- to 400-fold increase in mean lethal dose (LD50). The decrease in virulence produced by plasmids R68 and R68.45 was significantly greater than the decrease caused by the closely related plasmid RP1. All plasmids also led to an increased sensitivity of strain PAO1 to human serum bactericidal activity. Virulence and serum resistance of strain PAO1 were restored by curing of the entire plasmid R68.45 but not by deletions in the plasmid's transfer gene regions.

Background The non-conventional yeast Arxula adeninivorans uses 1-butanol as a carbon source and has recently attracted attention as a promising organism for 1-butanol production. Alcohol dehydrogenases (adhp) are important catalysts in 1-butanol metabolism, but only Aadh1p from Arxula has been characterized. This enzyme is involved in ethanol synthesis but has a low impact on 1-butanol degradation. Results In this study, we identified and characterized a second adhp from A. adeninivorans (Aadh2p). Compared to Saccharomyces cerevisiae ADHs’ (ScAdh) protein sequences it originates from the same ancestral node as ScAdh6p, 7p and 4p. It is also localized in the cytoplasm and uses NAD(H) as cofactor. The enzyme has its highest activity with medium chain-length alcohols and maximum activity with 1-butanol with the catalytic efficiency of the purified enzyme being 42 and 43,000 times higher than with ethanol and acetaldehyde, respectively. Arxula adeninivorans strain G1212/YRC102-AADH2, which expresses the AADH2 gene under the control of the strong constitutive TEF1 promoter was constructed. It achieved an ADH activity of up to 8000 U/L and 500 U/g dry cell weight (dcw) which is in contrast to the control strain G1212/YRC102 which had an ADH activity of up to 1400 U/L and 200 U/g dcw. Gene expression analysis showed that AADH2 derepression or induction using non-fermentable carbon-sources such as ethanol, pyruvate, glycerol or 1-butanol did occur. Compared to G1212/YRC102 AADH2 knock-out strain had a slower growth rate and lower 1-butanol consumption if 1-butanol was used as sole carbon source and AADH2-transformants did not grow at all in the same conditions. However, addition of the branched-chain amino acids leucine, isoleucine and valine allowed the transformants to use 1-butanol as carbon source. The addition of these amino acids to the control strain and Δaadh2 mutant cultures had the effect of accelerating 1-butanol consumption. Conclusions Our results confirm that Aadh2p plays a major role in A. adeninivorans 1-butanol metabolism. It is upregulated by up to 60-fold when the cells grow on 1-butanol, whereas only minor changes were found in the relative expression level for Aadh1p. Thus the constitutive overexpression of the AADH2 gene could be useful in the production of 1-butanol by A. adeninivorans, although it is likely that other ADHs will have to be knocked-out to prevent 1-butanol oxidation. Electronic supplementary material The online version of this article (doi:10.1186/s12934-016-0573-9) contains supplementary material, which is available to authorized users.

Web Usage Mining (WUM) aims to extract navigation usage patterns from Web server logs. Mining algorithms yield usage patterns, but finding the ones that constitute new and interesting knowledge in the domain remains a challenge. Typically, analysts have to deal with a huge volume of pattern, from which they have to retrieve the potentially interesting one and interpret what they reveal about the domain. In this paper, we discuss the filtering mechanisms of O3R, an environment supporting the retrieval and interpretation of sequential navigation patterns. All O3R functionality is based on the availability of the domain ontology, which dynamically provides meaning to URLs. The analyst uses ontology concepts to define filters, which can be applied according to two filtering mechanisms: equivalence and similarity.

Os avanços na tecnologia proporcionaram o aumento crescente na geração de dados e nos novos tipos de dados, tornando necessário estender os SGBDs para possibilitar armazenar, recuperar e organizar novos tipos de dados como imagens, vídeos e áudios, sendo estes conhecidos como dados complexos. Para as consultas em dados complexos, não é adequado comparar objetos utilizando as relações de Ordem e Identidade, sendo então a opção mais utilizada a comparação por similaridade. Dessa maneira, com a necessidade de desenvolver novos índices para as comparações baseadas em similaridade, surgiram os Métodos de Acesso Métricos (MAMs). Entre as diversas estratégias para indexar os dados, as baseadas em árvore se destacam por possibilitar um equilíbrio entre o tempo de construção do índice e a aceleração da consulta, sendo utilizada junto com a estratégia de árvore, uma estratégia para definir a região dos nós. Entre as diversas estratégias para definir regiões, o raio de cobertura está dentre as mais comumente utilizadas por flexibilizar a posição do objeto na estrutura, possibilitando o controle da ocupação dos nós e a redução no custo da construção da estrutura. Porém, esta estratégia possui o problema da sobreposição de nós, que aumenta o custo para obter as respostas exatas ao realizar as consultas por similaridade. Outra estratégia que não possui o problema da sobreposição, mas que sofre com o alto custo de construção, é a baseada no Diagrama de Voronoi. Buscando reduzir o problema da sobreposição de nós, aumentando o mínimo possível o custo da construção da árvore, neste projeto de mestrado foi proposto o MAM VD-Tree que busca acelerar as consultas por similaridade por meio da redução da sobreposição, obtida com reorganizações baseadas no Diagrama de Voronoi. Resultados experimentais mostraram que o método é capaz de acelerar consultas por similaridade e reduzir a sobreposição de nós na maioria dos casos, em comparação com seu principal competidor, o Slim-Tree. A melhora no tempo gasto ocorre devido ao método criar organizações melhores dos objetos na estrutura e reduzir a sobreposição dos nós, com o custo de criar mais nós para indexar os dados. Advances in the information technology have increased the amount of data generated daily and new types of data, making it necessary to extend DBMS to enable storing, retrieving, and organizing new types of data such as images, videos, and audio, known as complex data. It is not suitable for queries on complex data to compare objects using Order or Identity relations, so comparisons by similarity are the most employed option. With the necessity of developing new indices for comparisons based on similarity, many studies proposed several Metric Access Methods (MAMs). One of the most commonly used strategies to index complex data, tree-based strategies are commonly employed since they maintain a balance between the cost to create the index and the cost to execute the queries. Accordingly, together with the tree strategy, it is necessary to use a strategy to define the region of the nodes. Among the several strategies to define regions, the coverage radius strategy is commonly used to make the objects position in the structure more flexible, making it possible to control the occupation of nodes and reduce the cost of building the structure. However, this strategy has the problem of overlapping nodes, which increases the cost of getting the exact answers when performing similarity queries. Another strategy that does not have the overlap problem but suffers from the high construction cost is based on the Voronoi Diagram. Seeking to reduce the problem of overlapping nodes, increasing as little as possible the cost of constructing the tree, we propose here the VD-Tree MAM to speed up similarity queries by reducing the overlap between nodes, obtained with reorganizations based on the Voronoi Diagram. Experimental results showed that the method could speed up similarity queries with better distributions of the objects in the structure and reduce overlapping nodes in most cases, compared to its main competitor Slim-Tree, with the cost of requiring more nodes to index the data.

O custo do acesso exato a dados complexos tende a ser muito alto, do ponto de vista da carga de processamento computacional. Além disso, a operação de busca em dados multimídia não é efetuada realmente sobre os dados originais, mas sobre características extraídas desses dados, as quais os descrevem. Por exemplo, na busca por imagens similares utilizando-se histogramas de cor, realizando uma consulta exata, o que se obtém são as imagens cujos histogramas são exatamente os mais similares aos da imagem referenciada 11a consulta, mas isso não implica necessariamente que se obtenha as imagens que atendam exatamente a consulta efetuada, pois as imagens recuperadas podem ser muito diferentes quanto a forma, por exemplo. Portanto, em muitas aplicações que acessam dados complexos, a recuperação exata deixa de um requisito fundamental, podendo a exatidão das respostas ser trocada por um melhor desempenho Neste trabalho foram desenvolvidos algoritmos para recuperação aproximada do conjunto-resposta de consultas por similaridade em domínios métricos utilizando algoritmos genéticos. Neste trabalho, com a utilização de algoritmos genéticos, foram desenvolvidas técnicas de recuperação aproximada de dados cm domínio métrico, cujo refinamento das respostas obtidas é dependente do tempo de processamento disponível, definido pelo usuário. Os algoritmos desenvolvidos foram implementados utilizando a Slim-tree, mas outros métodos de acesso podem ser extendidos para utilizá-los também. Os algoritmos contemplam consultas por abrangência e vizinhos mais próximos, além de algumas variações. Os algoritmos desenvolvidos foram testados e validados com conjuntos de dados sintéticos e reais, e mostraram-se capazes de obter respostas aproximadas com boa precisão utilizando apenas uma fração do tempo exigido pela consulta exata. Os resultados obtidos nos experimentos mostram que é possível obter respostas com precisão superior a 90% utilizando apenas metade do tempo da consulta exata com o algoritmo convencional, e até 65% de precisão com apenas um quinto do tempo da consulta exata. I search process on complex domains for exact answer to a similarity query is an expensive process considering computational resources, such as memory and processing time requirements. However, when comparing multimédia dal,a, the comparison operations usually consider some properties of each daturn element, so exact queries involving this data return results tliat are exact regarding the properties compared, but not necessarily exact regarding the multimedia data itself. For example, searching for similar images regarding their colors return images whose color histogram are the most similar, but the retrieved images can be very different regarding for example the forni of the objects pietured. Therefore, for applications dealing with complex data types, trading exact answering with query time response can be a worth exchange. In this work we developed techniques based 011 genetic algorithms to allow retrieving approximate data indexed in a Metric Access Methods (MAMs) within a limited, user-defined, amount of time. For evaluation purposes, the algorithms were developed regarding the Slim-lrce, but the approximate query techniques developed in this work can be straightforwardly implemented on other MAMs. The algorithms can be used to perform nearest neighbor queries, range queries and some other variations. Svnthetic and real world datasets were used to evaluate the approximate algorithms, achieving good results in a fraetion of the time required to obtain the exact answer. The experimental results show that, allowing the algorithm to run during 50% of the exact query time, the precision of the approximate results is about 90%, and precision of 65% can be obtained consuming just 20% of the same exact query time.

International audience

International audience; Les nuées et le brouillard envahissent depuis longtemps nos représentations, et les techniques picturales, photographiques ou cinématographiques créent aujourd’hui encore des effets de confusion chez le spectateur et l’amènent à se confronter à l’opacité du visible. Mais la brume est aujourd’hui un opérateur plastique aussi bien dans les représentations bidimensionnelles que dans des installations où elle permet à un spectateur / acteur de faire l’expérience physique d’un lieu où les choses deviennent floues.Il s’agit d’analyser de manière croisée les enjeux contemporains de ces modes d’immersion (visuelle et/ou physique) du spectateur, dans les représentations brumeuses et dans des dispositifs plastiques à brouillard artificiel (Gerhard Richter, Olafur Eliasson, Ann-Veronica Janssens…). L’hypothèse est que le brouillard, qui place nos sens en alerte, permet aujourd’hui aux artistes de proposer au spectateur une perception altéritaire du monde ; cette dernière se fonde, paradoxalement, sur l’altération, dans une époque où le scintillement médiatique permanent finit par brouiller le sens des images évidentes et des signes injonctifs de tous ordres.-------Quand la brume se lève, l'espace diaphane s'opacifie ; elle perturbe notre environnement qui devient, littéralement, problématique. Se pose alors concrètement à nos sens la question de l'opacité et de la transparence du visible, et donc celle de notre compréhension du monde. Quand cette expérience d'un vide altéré et d'un réel brouillé se transpose dans celle de l'oeuvre plastique, il faut préciser ce qu'on entend par transparence et opacité. La représentation picturale symbolique ou naturaliste considérait la surface de la toile comme écran transparent qui ouvrait à une perception du monde fondée sur l'illusion et le décryptage. Mais en abandonnant le réalisme et la peinture d'idées au début du XXe siècle, l'art moderne ouvre une voie nouvelle : l'oeuvre devient opaque dans la mesure où elle opère de manière autonome dans le champ du réel, sans transparence ni référent obligatoire à l'extérieur d'elle-même. La peinture comme la sculpture acquièrent une autonomie : elles valorisent désormais davantage la visibilité de l'objet et la présence de la matière que l'identification d'une ressemblance. Cette conception moderne de l'oeuvre concerne bien sûr la forme de la brume comme opérateur plastique ; il ne s'agira donc pas seulement d'envisager ici la transparence et l'opacité qui caractérisent le brouillard en termes d'histoire et d'esthétique générale, mais également et surtout en termes phénoménologiques. Dans les arts visuels, la brume oscille désormais entre sa reconnaissance comme « vapeur colorée » dans une image et sa mise en oeuvre comme matériau réel opacifiant dans des dispositifs tridimensionnels. J'envisage donc le brouillard comme puissance scénographique qui établit des transactions et des passages dans le visible selon une dialectique transparence/opacité. Comment faisons-nous aujourd'hui l'expérience d'espaces plastiques brumeux, entre net et flou, entre réel et fiction ? Qu'il s'agisse d'un voile qui altère les repères en profondeur dans une oeuvre bidimensionnelle, d'un brouillard artificiel qui nous désoriente dans l'espace d'exposition ou encore d'une brume en mouvement qui envahit le paysage dans un dispositif in situ, les artistes proposent au spectateur une expérience sensorielle sur l'émergence, la transformation et la disparition des formes ou des choses.

International audience

Decision-tree induction is one of the most employed methods to extract knowledge from data. There are several distinct strategies for inducing decision trees from data, each one presenting advantages and disadvantages according to its corresponding inductive bias. These strategies have been continuously improved by researchers over the last 40 years. This thesis, following recent breakthroughs in the automatic design of machine learning algorithms, proposes to automatically generate decision-tree induction algorithms. Our proposed approach, namely HEAD-DT, is based on the evolutionary algorithms paradigm, which improves solutions based on metaphors of biological processes. HEAD-DT works over several manually-designed decision-tree components and combines the most suitable components for the task at hand. It can operate according to two different frameworks: i) evolving algorithms tailored to one single data set (specific framework); and ii) evolving algorithms from multiple data sets (general framework). The specific framework aims at generating one decision-tree algorithm per data set, so the resulting algorithm does not need to generalise beyond its target data set. The general framework has a more ambitious goal, which is to generate a single decision-tree algorithm capable of being effectively applied to several data sets. The specific framework is tested over 20 UCI data sets, and results show that HEAD-DTs specific algorithms outperform algorithms like CART and C4.5 with statistical significance. The general framework, in turn, is executed under two different scenarios: i) designing a domain-specific algorithm; and ii) designing a robust domain-free algorithm. The first scenario is tested over 35 microarray gene expression data sets, and results show that HEAD-DTs algorithms consistently outperform C4.5 and CART in different experimental configurations. The second scenario is tested over 67 UCI data sets, and HEAD-DTs algorithms were shown to be competitive with C4.5 and CART. Nevertheless, we show that HEAD-DT is prone to a special case of overfitting when it is executed under the second scenario of the general framework, and we point to possible alternatives for solving this problem. Finally, we perform an extensive experiment for evaluating the best single-objective fitness function for HEAD-DT, combining 5 classification performance measures with three aggregation schemes. We evaluate the 15 fitness functions in 67 UCI data sets, and the best of them are employed to generate algorithms tailored to balanced and imbalanced data. Results show that the automatically-designed algorithms outperform CART and C4.5 with statistical significance, indicating that HEAD-DT is also capable of generating custom algorithms for data with a particular kind of statistical profile Árvores de decisão são amplamente utilizadas como estratégia para extração de conhecimento de dados. Existem muitas estratégias diferentes para indução de árvores de decisão, cada qual com suas vantagens e desvantagens tendo em vista seu bias indutivo. Tais estratégias têm sido continuamente melhoradas por pesquisadores nos últimos 40 anos. Esta tese, em sintonia com recentes descobertas no campo de projeto automático de algoritmos de aprendizado de máquina, propõe a geração automática de algoritmos de indução de árvores de decisão. A abordagem proposta, chamada de HEAD-DT, é baseada no paradigma de algoritmos evolutivos. HEAD-DT evolui componentes de árvores de decisão que foram manualmente codificados e os combina da forma mais adequada ao problema em questão. HEAD-DT funciona conforme dois diferentes frameworks: i) evolução de algoritmos customizados para uma única base de dados (framework específico); e ii) evolução de algoritmos a partir de múltiplas bases (framework geral). O framework específico tem por objetivo gerar um algoritmo por base de dados, de forma que o algoritmo projetado não necessite de poder de generalização que vá além da base alvo. O framework geral tem um objetivo mais ambicioso: gerar um único algoritmo capaz de ser efetivamente executado em várias bases de dados. O framework específico é testado em 20 bases públicas da UCI, e os resultados mostram que os algoritmos específicos gerados por HEAD-DT apresentam desempenho preditivo significativamente melhor do que algoritmos como CART e C4.5. O framework geral é executado em dois cenários diferentes: i) projeto de algoritmo específico a um domínio de aplicação; e ii) projeto de um algoritmo livre-de-domínio, robusto a bases distintas. O primeiro cenário é testado em 35 bases de expressão gênica, e os resultados mostram que o algoritmo gerado por HEAD-DT consistentemente supera CART e C4.5 em diferentes configurações experimentais. O segundo cenário é testado em 67 bases de dados da UCI, e os resultados mostram que o algoritmo gerado por HEAD-DT é competitivo com CART e C4.5. No entanto, é mostrado que HEAD-DT é vulnerável a um caso particular de overfitting quando executado sobre o segundo cenário do framework geral, e indica-se assim possíveis soluções para tal problema. Por fim, é realizado uma análise detalhada para avaliação de diferentes funções de fitness de HEAD-DT, onde 5 medidas de desempenho são combinadas com três esquemas de agregação. As 15 versões são avaliadas em 67 bases da UCI e as melhores versões são utilizadas para geração de algoritmos customizados para bases balanceadas e desbalanceadas. Os resultados mostram que os algoritmos gerados por HEAD-DT apresentam desempenho preditivo significativamente melhor que CART e C4.5, em uma clara indicação que HEAD-DT também é capaz de gerar algoritmos customizados para certo perfil estatístico dos dados de classificação

International audience; La Fête de la Fédération a-t-elle été une belle journée ? Tout est question d’écriture. La valeur de symbole ou de signe que prennent la pluie et les rares rayons de soleil, le sens donné aux réactions qu’ils provoquent et les sentiments qu’ils suscitent sont l’occasion de conflits d’interprétation et d’usages rhétoriques divergents. Si les jugements de valeur et les commentaires explicites sont nombreux, c’est aussi par la sélection et la mise en ordre opérées que chaque texte fait sens.Quelle influence les discours météorologiques et leurs usages métaphoriques ont-il eu les uns sur les autres à l’occasion de cet événement climatique et politique ? La politisation des discours météorologiques est évidente. La pluie se charge de sens dans le contexte solennel de la Fédération. Dans un mouvement inverse, les phénomènes climatiques réels modifient les usages métaphoriques. Si la pluie peut toujours signifier les désordres et les peines et le soleil dire la vérité et le bonheur, les textes qui rendent compte de la Fédération s’en tiennent rarement à cette valeur symbolique ou à une signification intrinsèque de la pluie et du soleil considérés comme des signes divins. La pluie du 14 juillet 1790 éloigne la plupart des commentateurs du sacré. Avec l’expression « dieu est aristocrate », les signes sont tournés en dérision, que ce soit dans la satire des contre-révolutionnaires ou dans les plaisanteries des fédérés. Ce qui compte en revanche, c’est l’expérience des intempéries, à travers les réactions qu’elles provoquent et les sentiments qu’elles suscitent. Même dans un contexte symbolique aussi chargé que celui de la Fédération, la pluie et le beau temps sont des objets d’observations extérieures et intérieures. Épreuve politique et collective, expérience sensible et individuelle, les intempéries de la Fédération sont l’occasion d’appliquer le thermomètre de l’opinion et le baromètre de l’âme.