1,720,974 research outputs found
Identification structurelle
No full textDe nombreuses techniques mathématiques utilisées en robotique reposent sur l'identification de paramètres ou sur la construction d'un modèle de type boîte noire (réseaux de neurones par exemple). Dans le premier cas on se donne une équation de mesure dont on ignore certains paramètres, mais pour laquelle on dispose de mesures experimentales du phénomène qu'on cherche à modéliser. Le problème revient alors à trouver les valeurs numériques des paramètres inconnus de l'équation pour obtenir un modèle du système. Cela revient en général à conduire un certain nombre d'expériences puis à faire de la minimisation au sens des moindres carrés pour trouver les paramètres qui expliquent au mieux les mesures obtenues. Dans le deuxième cas on utilise une technique d'approximation universelle permettant de modéliser la réponse du système. Pour cela on corrige l'ensemble des paramètres de l'approximateur à l'aide d'un algorithme adaptatif et d'un ensemble d'exemples. On dispose donc actuellement de deux grandes classes de méthodes : l'une faisant appel à de fortes connaissances préalables (la connaissance de l'équation de mesure) et l'autre ne faisant appel à aucune connaissance préalable. L'objet de cette thèse est de proposer une méthode intermédiaire: l'identification structurelle. Dans ce cadre on ne connaît plus la forme paramétrique de l'équation de mesure mais des informations a-priori sur sa forme générale. Par exemple, on sait que l'équation de mesure est formée d'un polynôme de fonctions quelconques d'une seule variable. Nous montrons qu'il est possible d'inférer cette équation de mesure dès lors que l'on choisit un protocole expérimental approprié et que l'on dispose d'un approximateur universel pour les fonctions d'une seule variable. L'ensemble des polynômes de fonctions trigonométiques multi-variables rentre dans le cadre juste évoqué. On peut donc appliquer cette méthode à de nombreux problèmes trouvés en robotique. On peut par exemple identifier le modèle géométrique d'un bras manipulateur ou trouver l'expression de la jacobienne reliant les mouvements d'un bras aux mouvements d'indices visuels dans une image vidéo. Le modèle fonctionnel obtenu peut être utilisé pour commander le système. C'est ainsi que nous avons réalisé un asservissement visuel avec cette méthode
Uma rede neural para o reconhecimento de padrões codificados em sequências
Full text linkEste trabalho apresenta um modelo de rede neural voltado ao tratamento de informagbes codificadas em sequencias, tendo em vista que esta classe de informações nao tem um tratamento adequado nos modelos convencionais. Isso decorre da caracteristica destes modelos convencionais de manipular isoladamente as celulas de informacao apresentadas como entrada, sem Integra-las entre si. O modelo utiliza paradigmas a mecanismos conhecidos, tais coma a regra de HEBB, o modelo de Energia de Hopfield e o paradigma de organizacao em camadas, compondo-os com novas ideias e mecanismos direcionados para o tratamento de sequencias, em um sistema exploratório, extraindo com isso novas propriedades nao existentes em modelos tradicionais. Os novas mecanismos propostos permitem uma integragao entre entradas a rede e o contexto no qual elas sao apresentadas, para que com isso se forme uma Unica representacAo interna para Coda uma seqUencia de entradas. Todo o trabalho de validaco do modelo foi baseado em simulac6es, para as quaffs foi desenvolvido um ambiente em estacao de trabalho, dotado de interface grafica que permite o acompanhamento visual do funcionamento da rede. Para viabilizar a validacao do modelo por meio das simulac6es, tendo em vista os limites computacionais dos recursos disponiveis, foi proposto e utilizado um sistema de codificacao de informacbes ficticias simplicadas baseadas na fala, mais especificamente na organizacao fonetica. O sistema de codificação, embora simplificado, incorpora as mais importantes caracteristicas da codificação de informacbes realizada na fala, pelo menos pelo ponto de vista de seu reconhecimento por mein de redes neurais.This work presents a Neural Network model to process sequence information, since this information class does not have a reasonable treatment in the conventional models. This is due to this models features that manipulate incoming information cells individually, without integrating them. The model uses already known mechanisms and paradigms, like the HEBB's rule, the Hopfield's Energy Model and the layer organization paradigm, added with new ideas and mechanisms for the sequence handling in a exploratory system, so that it extracts new properties not found in traditional models. The proposed new mechanisms allow the integration between network entries and context , in order to generate a unique internal representation. The model was validated through simulations. A workstation based environment was designed and implemented to support them. It incorporates a graphical interface that permits the network behavior visualization. In order to enable the model's validation through simulations and considering the computational limits of the available resources, a codification system was proposed to generate simplified ficticious and speech based informations. Although simplified, this codification system incorporates the most important features of the information codification that occur in the speech, at least from its neural network based recognition point of view
Going Beyond Counting First Authors in Author Co-citation Analysis
Full text linkThe present study examines one of the fundamental aspects of author co-citation analysis (ACA) - the way co-citation
counts are defined. Co-citation counting provides the data on which all subsequent statistical analyses and mappings
are based, and we compare ACA results based on two different types of co-citation counting - the traditional type that
only counts the first one among a cited work's authors on the one hand and a non-traditional type that takes into
account the first 5 authors of a cited work on the other hand. Results indicate that the picture produced through this non-traditional author co-citation counting contains more coherent author groups and is therefore considerably clearer. However, this picture represents fewer specialties in the research field being studied than that produced through the traditional first-author co-citation counting when the same number of top-ranked authors is selected and analyzed. Reasons for these effects are discussed
Variations on the Author
Full text link“Variations on the Author” discusses two of Eduardo Coutinho’s recent films (Um Dia na Vida, from 2010, and Últimas Conversas, posthumously released in 2015) and their contribution to the general question of documentary authorship. The director’s filmography is characterized by a consistent yet self-effacing form of authorial self-inscription: Coutinho often features as an interviewer that rather than express opinions propels discourses; an interviewer that is good at listening. This mode of self-inscription characterizes him as an author who is not expressive but who is nonetheless markedly present on the screen. In Um Dia na Vida, however, Coutinho is completely absent form the image, while Últimas Conversas, on the contrary, includes a confessional prologue that moves the director from the margins to the center of his films. This article examines the ways in which these works stand out in the filmography of a director who offers new insights into the notion of cinematic authorship
Appropriate Similarity Measures for Author Cocitation Analysis
Full text linkWe provide a number of new insights into the methodological discussion about author cocitation analysis. We first argue that the use of the Pearson correlation for measuring the similarity between authors’ cocitation profiles is not very satisfactory. We then discuss what kind of similarity measures may be used as an alternative to the Pearson correlation. We consider three similarity measures in particular. One is the well-known cosine. The other two similarity measures have not been used before in the bibliometric literature. Finally, we show by means of an example that our findings have a high practical relevance.information science;Pearson correlation;cosine;similarity measure;author cocitation analysis
Dispelling the Myths Behind First-author Citation Counts
Full text linkWe conducted a full-scale evaluative citation analysis study of scholars in the XML research field to explore just how different from each other author rankings resulting from different citation counting methods actually are, and to demonstrate the capability of emerging data and tools on the Web in supporting more realistic citation counting methods. Our results contest some common arguments for the continued
use of first-author citation counts in the evaluation of scholars, such as high correlations between author rankings by first-author citation counts and other citation
counting methods, and high costs of using more realistic citation counting methods that are not well-supported by the ISI databases. It is argued that increasingly available digital full text research papers make it possible for citation analysis studies to go beyond what the ISI databases have directly supported and to employ more
sophisticated methods
koamabayili/VECTRON-author-checklist: VECTRON author checklist
No full textWe have done our best to complete the author checklist relating to the use of animals in the hut study. Note that the objective for the hut study was to evaluate the IRS treatment applications for residual efficacy against Anopheles mosquitoes, including the local An. coluzzii mosquito population. Cows were only used to attract mosquitoes into the huts and no tests were carried out directly on the cows. The author checklist is intended for use with studies where experiments are carried out on animals, which is why we have had such difficulty in completing this for the hut study, as many of the questions do not relate to how the cows were used
Identificando evasão fiscal em empresas de fachada e em créditos ilegais de ICMS
Full text linkCompanies that issue tax documents to defraud the tax authorities with the transfer of credits of Brazil’s state value-added tax (ICMS) without the movement of goods cause financial losses to the government and, therefore, to society as a whole. Several initiatives to combat tax fraud have successfully used data analysis and Machine Learning techniques. This work sought to investigate the use of these techniques in identifying a specific practice of tax fraud, practiced by shell companies, formed exclusively to issue non-due ICMS credits, the tax on operations related to the circulation of goods, and the provision of interstate, intercity, and communication services. Based on document analysis and consultation with auditors and specialists, typologies and variables relevant to identifying tax evasion events carried out by shell companies were identified. Around these variables, data from the Finance Department of the Federal District were collected and prepared. With this data, it was possible to explore the use of predictive models based on Machine Learning capable of pointing out potentially fraudulent behavior. The good results obtained by these models demonstrate their potential as part of systematic monitoring and fiscal audits by tax authorities.Las empresas que emiten documentos tributarios para defraudar al fisco con la transferencia de crédito del ICMS (impuesto a las operaciones relacionadas con la circulación. de bienes y de prestación de servicios interestatales, interurbanos y de comunicaciones) sin movimiento de mercancías causan danos al erario público y, por ende, a la sociedad en su conjunto. Varias iniciativas para combatir el fraude fiscal han utilizado con éxito técnicas de análisis de datos y aprendizaje automático. Este trabajo busco investigar el uso de estas técnicas en la identificación de una práctica especifica de fraude fiscal, practicada por empresas conocidas popularmente como ‘empresas factureras’, constituidas exclusivamente para emitir créditos no vencidos del ICMS. A partir del análisis documental y la consulta a auditores y especialistas, se identificaron tipologías y variables relevantes para la identificación de eventos de evasión fiscal realizados por empresas factureras. En torno a estas variables se recolectaron y prepararon datos desde la Secretaria de Hacienda del Distrito Federal. Con estos datos fue posible explorar el uso de modelos predictivos basados en machine learning capaces de señalar comportamientos potencialmente fraudulentos. Los buenos resultados obtenidos por estos modelos demuestran su potencial como parte de un seguimiento sistemático y auditorias fiscales por parte de las autoridades tributarias.Empresas que emitem documentos fiscais para fraudar o fisco com a transferência de crédito do ICMS sem a circulação de mercadorias causam prejuízo ao erário público e, por conseguinte, a sociedade. Diversas iniciativas de combate a fraudes fiscais têm utilizado, com sucesso, técnicas de análise de dados e aprendizagem de máquina. Este trabalho buscou investigar o uso dessas técnicas na identificação de uma prática específica de fraude fiscal realizada por empresas popularmente conhecidas como “empresas noteiras”, que formadas exclusivamente para emitir créditos não devidos de ICMS, imposto sobre operações relativas à circulação de mercadorias e sobre prestações de serviços de transporte interestadual, intermunicipal e de comunicação. Com base na análise documental e em consulta com auditores e especialistas, foram identificadas tipologias e variáveis relevantes na determinação de eventos de sonegação fiscal realizados pelas empresas noteiras. Em torno dessas variáveis, procedeu-se a coleta e a preparação de dados provenientes da Secretaria de Fazenda do Distrito Federal. Com esses dados, foi possível explorar o uso de modelos preditivos baseados em aprendizagem de máquina capazes de apontar comportamentos potencialmente fraudulentos. Os bons resultados obtidos por esses modelos demonstram seu potencial como parte de uma sistemática de monitoramento e auditorias fiscais realizadas pelos órgãos fazendários
- …
