Improving the accuracy of the optimum-path forest supervised classifier for large datasets

In this work, a new approach for supervised pattern recognition is presented which improves the learning algorithm of the Optimum-Path Forest classifier (OPF), centered on detection and elimination of outliers in the training set. Identification of outliers is based on a penalty computed for each sample in the training set from the corresponding number of imputable false positive and false negative classification of samples. This approach enhances the accuracy of OPF while still gaining in classification time, at the expense of a slight increase in training time. © 2010 Springer-Verlag.

The Role of Technical Assistance in World Trade Organization (WTO) Trade Facilitation Negotiations

This issue of the Bulletin presents a brief review of trade facilitation negotiations from the specific viewpoint of technical assistance, capacity-building and special and differential treatment (SDT). Trade facilitation negotiations have focused on clarifying three articles of the General Agreement on Tariffs and Trade (GATT): article V (freedom of transit); article VIII (fees and formalities connected with importation and exportation); and article X (publication and administration of trade regulations). Although the stage of text-based negotiations has not yet been reached, the process of receiving and analysing proposals has made it possible to identify the aspects most urgently in need of correcting by means of trade facilitation measures. Consideration has been given to several formulas for the implementation of those measures and to how the associated technical assistance should be organized.

As cartas de Chico Xavier: uma análise semiótica

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

Ferramentas e recursos livres para reconhecimento e síntese de voz em português brasileiro

Sistemas de reconhecimento e síntese de voz são constituídos por módulos que dependem da língua e, enquanto existem muitos recursos públicos para alguns idiomas (p.e. Inglês e Japonês), os recursos para Português Brasileiro (PB) ainda são escassos. Outro aspecto é que, para um grande número de tarefas, a taxa de erro dos sistemas de reconhecimento de voz atuais ainda é elevada, quando comparada à obtida por seres humanos. Assim, apesar do sucesso das cadeias escondidas de Markov (HMM), é necessária a pesquisa por novos métodos. Este trabalho tem como motivação esses dois fatos e se divide em duas partes. A primeira descreve o desenvolvimento de recursos e ferramentas livres para reconhecimento e síntese de voz em PB, consistindo de bases de dados de áudio e texto, um dicionário fonético, um conversor grafema-fone, um separador silábico e modelos acústico e de linguagem. Todos os recursos construídos encontram-se publicamente disponíveis e, junto com uma interface de programação proposta, têm sido usados para o desenvolvimento de várias novas aplicações em tempo-real, incluindo um módulo de reconhecimento de voz para a suíte de aplicativos para escritório OpenOffice.org. São apresentados testes de desempenho dos sistemas desenvolvidos. Os recursos aqui produzidos e disponibilizados facilitam a adoção da tecnologia de voz para PB por outros grupos de pesquisa, desenvolvedores e pela indústria. A segunda parte do trabalho apresenta um novo método para reavaliar (rescoring) o resultado do reconhecimento baseado em HMMs, o qual é organizado em uma estrutura de dados do tipo lattice. Mais especificamente, o sistema utiliza classificadores discriminativos que buscam diminuir a confusão entre pares de fones. Para cada um desses problemas binários, são usadas técnicas de seleção automática de parâmetros para escolher a representaçãao paramétrica mais adequada para o problema em questão.

Habilidades sociais em portadores de anomalia da diferenciação sexual

O termo Habilidades Sociais (HS) é usado para designar um conjunto de repertórios comportamentais que envolvem interações sociais satisfatórias. A maioria dos estudos publicados sobre essa temática inclui a participação de crianças e adolescentes, abordando situações em ambiente escolar. Poucos são os estudos voltados para a área de saúde, e, até o momento, não foram localizadas pesquisas sobre HS realizadas com indivíduos portadores de Anomalia da Diferenciação Sexual (ADS). Esta anomalia se caracteriza pela malformação na genitália e/ou pela disfunção das gônadas, ocasionando características sexuais secundárias não correspondentes ao sexo de criação. Estudos clínicos realizados com indivíduos portadores de ADS têm destacado a ocorrência de déficits em habilidades sociais nesses indivíduos, caracterizados pela fuga-esquiva de situações sociais aversivas. Neste trabalho, pretendeu-se caracterizar o repertório comportamental correspondente a HS em indivíduos com ADS atendidos no ambulatório de um programa especializado de um hospital da rede pública de Belém, por meio de dois estudos complementares. No primeiro, foi realizado um estudo com delineamento transversal, com o objetivo de caracterizar comportamentos correspondentes a HS em indivíduos com ADS. Participaram 9 adultos com mais de seis meses em tratamento. Foram utilizados: Roteiro de Entrevista, Protocolo para análise de prontuário e Inventário de Habilidades Sociais - IHS. Os resultados sugerem que os participantes apresentam déficit de habilidades sociais em todos os fatores do IHS em menor ou maior grau, de maneira que poderiam se beneficiar com um treinamento de habilidades sociais como parte do tratamento, por se tratar de uma técnica que visa à superação e/ou redução dos déficits por eles apresentados. No segundo, foi realizado um estudo com delineamento de sujeito único, com o objetivo de verificar os efeitos do uso de treino em automonitoramento na instalação de comportamentos correspondentes a HS. Participou uma adulta com diagnóstico de ADS selecionada dentre os que participaram do Estudo 1. O procedimento de intervenção ocorreu por meio de entrevistas semanais, de acordo com as seguintes etapas: (1) Contrato: assinatura do Termo de Consentimento Livre e Esclarecido e agendamento de entrevista; (2) Avaliação: levantamento da linha de base dos comportamentos correspondentes à HS, elaboração da hierarquia dos comportamentos que indicaram déficit segundo as normas de análise do IHS e treino em registro de automonitoração; (3) Intervenção: leitura de um texto sobre habilidades sociais, apresentação da lista de direitos humanos básicos, aplicação do Questionário Construcional de Goldiamond adaptado e treino em registro de automonitoramento; (4) Re-avaliação das HS: reaplicação do IHS; (5) Follow-up: avaliação da manutenção dos ganhos obtidos com o estudo; e (6) Encerramento: entrevista devolutiva para explanação dos resultados do estudo ao participante. Os resultados sugerem que a intervenção promoveu o desenvolvimento de HS na participante, haja vista a interpretação da evolução da participante apresentada de acordo com a variação do Escore Z e da representação gráfica de significância clínica e mudança confiável.

Construction and validation of an instrument to assess the reading comprehension of students from the third to the fifth grades of elementary school

Objetivo:Este estudo teve como objetivo construir e validar um instrumento de avaliação da compreensão de leitura a fim de caracterizar o perfil de leitura e detectar dificuldades de compreensão em escolares do terceiro ao quinto ano do Ensino Fundamental.Métodos:Participaram 378 escolares divididos em trés grupos para avaliação da compreensão de proposições literais e inferenciais de micro e macroestruturas de dois textos expositivos e dois textos narrativos por meio de questões de múltipla escolha.Resultados:Os dados analisados estatisticamente indicaram valores do teste alfa de Cronbach apresentando consisténcia interna nos quatros textos aplicados para os trés grupos.Conclusão:Foi possível constatar que os escolares apresentaram menor número de erros com o aumento da escolarização e que cada tipo de texto apresentou uma dificuldade específica para os escolares.

Preliminary diagnosis of ophtalmological diseases through machine learning techniques

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

Complex networks analysis of language complexity

Methods from statistical physics, such as those involving complex networks, have been increasingly used in the quantitative analysis of linguistic phenomena. In this paper, we represented pieces of text with different levels of simplification in co-occurrence networks and found that topological regularity correlated negatively with textual complexity. Furthermore, in less complex texts the distance between concepts, represented as nodes, tended to decrease. The complex networks metrics were treated with multivariate pattern recognition techniques, which allowed us to distinguish between original texts and their simplified versions. For each original text, two simplified versions were generated manually with increasing number of simplification operations. As expected, distinction was easier for the strongly simplified versions, where the most relevant metrics were node strength, shortest paths and diversity. Also, the discrimination of complex texts was improved with higher hierarchical network metrics, thus pointing to the usefulness of considering wider contexts around the concepts. Though the accuracy rate in the distinction was not as high as in methods using deep linguistic knowledge, the complex network approach is still useful for a rapid screening of texts whenever assessing complexity is essential to guarantee accessibility to readers with limited reading ability. Copyright (c) EPLA, 2012

Influence of the magnetization damping on dynamic hysteresis loops in single domain particles

This article reports on the influence of the magnetization damping on dynamic hysteresis loops in single-domain particles with uniaxial anisotropy. The approach is based on the Neel-Brown theory and the hierarchy of differential recurrence relations, which follow from averaging over the realizations of the stochastic Landau-Lifshitz equation. A new method of solution is proposed, where the resulting system of differential equations is solved directly using optimized algorithms to explore its sparsity. All parameters involved in uniaxial systems are treated in detail, with particular attention given to the frequency dependence. It is shown that in the ferromagnetic resonance region, novel phenomena are observed for even moderately low values of the damping. The hysteresis loops assume remarkably unusual shapes, which are also followed by a pronounced reduction of their heights. Also demonstrated is that these features remain for randomly oriented ensembles and, moreover, are approximately independent of temperature and particle size. (C) 2012 American Institute of Physics. [doi:10.1063/1.3684629]

Seeing beyond reading: a survey on visual text analytics

We review recent visualization techniques aimed at supporting tasks that require the analysis of text documents, from approaches targeted at visually summarizing the relevant content of a single document to those aimed at assisting exploratory investigation of whole collections of documents.Techniques are organized considering their target input materialeither single texts or collections of textsand their focus, which may be at displaying content, emphasizing relevant relationships, highlighting the temporal evolution of a document or collection, or helping users to handle results from a query posed to a search engine.We describe the approaches adopted by distinct techniques and briefly review the strategies they employ to obtain meaningful text models, discuss how they extract the information required to produce representative visualizations, the tasks they intend to support and the interaction issues involved, and strengths and limitations. Finally, we show a summary of techniques, highlighting their goals and distinguishing characteristics. We also briefly discuss some open problems and research directions in the fields of visual text mining and text analytics.

Identification of literary movements using complex networks to represent texts

The use of statistical methods to analyze large databases of text has been useful in unveiling patterns of human behavior and establishing historical links between cultures and languages. In this study, we identified literary movements by treating books published from 1590 to 1922 as complex networks, whose metrics were analyzed with multivariate techniques to generate six clusters of books. The latter correspond to time periods coinciding with relevant literary movements over the last five centuries. The most important factor contributing to the distinctions between different literary styles was the average shortest path length, in particular the asymmetry of its distribution. Furthermore, over time there has emerged a trend toward larger average shortest path lengths, which is correlated with increased syntactic complexity, and a more uniform use of the words reflected in a smaller power-law coefficient for the distribution of word frequency. Changes in literary style were also found to be driven by opposition to earlier writing styles, as revealed by the analysis performed with geometrical concepts. The approaches adopted here are generic and may be extended to analyze a number of features of languages and cultures.

Huckebein is part of a combinatorial repression code in the anterior blastoderm

The hierarchy of the segmentation cascade responsible for establishing the Drosophila body plan is composed by gap, pair-rule and segment polarity genes. However, no pair-rule stripes are formed in the anterior regions of the embryo. This lack of stripe formation, as well as other evidence from the literature that is further investigated here, led us to the hypothesis that anterior gap genes might be involved in a combinatorial mechanism responsible for repressing the cis-regulatory modules (CRMs) of hairy (h), even-skipped (eve), runt (run), and fushi-tarazu (ftz) anterior-most stripes. In this study, we investigated huckebein (hkb), which has a gap expression domain at the anterior tip of the embryo. Using genetic methods we were able to detect deviations from the wild-type patterns of the anterior-most pair-rule stripes in different genetic backgrounds, which were consistent with Hkb-mediated repression. Moreover, we developed an image processing tool that, for the most part, confirmed our assumptions. Using an hkb misexpression system, we further detected specific repression on anterior stripes. Furthermore, bioinformatics analysis predicted an increased significance of binding site clusters in the CRMs of h 1, eve 1, run 1 and ftz 1 when Hkb was incorporated in the analysis, indicating that Hkb plays a direct role in these CRMs. We further discuss that Hkb and Slp1, which is the other previously identified common repressor of anterior stripes, might participate in a combinatorial repression mechanism controlling stripe CRMs in the anterior parts of the embryo and define the borders of these anterior stripes. (C) 2011 Elsevier Inc. All rights reserved.

Adaptation and Constraint in the Plant Reproductive Phase

Conservatism is a central theme of organismic evolution. Related species share characteristics due to their common ancestry. Some concern have been raised among evolutionary biologists, whether such conservatism is an expression of natural selection or of a constrained ability to adapt. This thesis explores adaptations and constraints within the plant reproductive phase, particularly in relation to the evolution of fleshy fruit types (berries, drupes, etc.) and the seasonal timing of flowering and fruiting. The different studies were arranged along a hierarchy of scale, with general data sets sampled among seed plants at the global scale, through more specific analyses of character evolution within the genus Rhamnus s.l. L. (Rhamnaceae), to descriptive and experimental field studies in a local population of Frangula alnus (Rhamnaceae). Apart from the field study, this thesis is mainly based on comparative methods explicitly incorporating phylogenetic relationships. The comparative study of Rhamnus s.l. species included the reconstruction of phylogenetic hypotheses based on DNA sequences. Among geographically overlapping sister clades, biotic pollination was not correlated with higher species richness when compared to wind pollinated plants. Among woody plants, clades characterized by fleshy fruit types were more species rich than their dry-fruited sister clades, suggesting that the fleshy fruit is a key innovation in woody habitats. Moreover, evolution of fleshy fruits was correlated with a change to more closed (darker) habitats. An independent contrast study within Rhamnus s.l. documented allometric relations between plant and fruit size. As a phylogenetic constraint, allometric effects must be considered weak or non-existent, though, as they did not prevail among different subclades within Rhamnus s.l. Fruit size was correlated with seed size and seed number in F. alnus. This thesis suggests that frugivore selection on fleshy fruit may be important by constraining the upper limits of fruit size, when a plant lineage is colonizing (darker) habitats where larger seed size is adaptive. Phenological correlations with fruit set, dispersal, and seed size in F. alnus, suggested that the evolution of reproductive phenology is constrained by trade-offs and partial interdependences between flowering, fruiting, dispersal, and recruitment phases. Phylogenetic constraints on the evolution of phenology were indicated by a lack of correlation between flowering time and seasonal length within Rhamnus cathartica and F. alnus, respectively. On the other hand, flowering time was correlated with seasonal length among Rhamnus s.l. species. Phenological differences between biotically and wind pollinated angiosperms also suggested adaptive change in reproductive phenology.

Is there a link between the type of habitat and the patterns of abundance of holothurians in shallow rocky reefs?

[EN] The presence of a mosaic of habitats, largely determined by sea urchin grazing, across shallow rocky reefs may potentially influence in differences in the distribution patterns of invertebrates. The aim of this paper was to assess, using a correlative approach, whether the type of habitat influences the abundance patterns of holothurians in the eastern Atlantic. We hypothesized that abundances of large (> 10 cm) holothurians varied among four types of habitat (3 vegetated habitats with low abundances of the sea urchin D. antillarum versus ?barrens? with hyperabundances of sea urchins), and that these differences were consistent at a hierarchy of spatial scales, including two islands and several replicated sites within each type of habitat and island. Three species of large holothurians were found, accounting for a total of 300 specimens. We found remarkable differences in abundances of holothurians between the ?barrens? and the three vegetated habitats. This pattern was strongest for the numerically dominant species, Holothuria sanctorii. Total abundances of holothurians were between 5 ? 46 times more abundant in ?barrens? compared with the vegetated habitats. Inter-habitat differences were species-specific with some inconsistent patterns from one island to the other. The total abundances of holothurians tended to increase with the abundance of sea urchins within ?barrens?. Our study suggests that there may be a link, at least for the dominant species Holothuria sanctorii, between the distribution and abundances of large holothurians and the habitat across shallow-waters of the eastern Atlantic.

Expressiveness of Concurrent Languages

The aim of this thesis is to go through different approaches for proving expressiveness properties in several concurrent languages. We analyse four different calculi exploiting for each one a different technique. We begin with the analysis of a synchronous language, we explore the expressiveness of a fragment of CCS! (a variant of Milner's CCS where replication is considered instead of recursion) w.r.t. the existence of faithful encodings (i.e. encodings that respect the behaviour of the encoded model without introducing unnecessary computations) of models of computability strictly less expressive than Turing Machines. Namely, grammars of types 1,2 and 3 in the Chomsky Hierarchy. We then move to asynchronous languages and we study full abstraction for two Linda-like languages. Linda can be considered as the asynchronous version of CCS plus a shared memory (a multiset of elements) that is used for storing messages. After having defined a denotational semantics based on traces, we obtain fully abstract semantics for both languages by using suitable abstractions in order to identify different traces which do not correspond to different behaviours. Since the ability of one of the two variants considered of recognising multiple occurrences of messages in the store (which accounts for an increase of expressiveness) reflects in a less complex abstraction, we then study other languages where multiplicity plays a fundamental role. We consider the language CHR (Constraint Handling Rules) a language which uses multi-headed (guarded) rules. We prove that multiple heads augment the expressive power of the language. Indeed we show that if we restrict to rules where the head contains at most n atoms we could generate a hierarchy of languages with increasing expressiveness (i.e. the CHR language allowing at most n atoms in the heads is more expressive than the language allowing at most m atoms, with mof the rewriting rules, several dialects of the calculus can be obtained. We analyse the expressive power of some of these dialects by focusing on decidability and undecidability for problems like reachability and coverability.