This diploma thesis researchs the collocability of the german adjectives and gives some suggestions for the presentation of the collocability in the microstructure of the adjectival entries in the German-Czech translatory dictionary. The diploma thesis is a secondary product of the project of The Large German-Czech Academic Dictionary which has been realized in the Institute for German studies of the Phaculty of Arts of Charles University in Prague since 2000. The lexicographs use the methods of the corpus linguistics, especially a combination of the concepts corpus-based and corpus-driven. Analogous methods were used in this diploma thesis. By means of the Cooccurrence Data Bank (CCDB) of the Institute for the German language in Mannheim the most frequent cooccurrence partners of an adjectiv were found, concrete dictionary examples were found out in the corpus DeReKo of the IDS Mannheim using COSMAS II or on the Internet using the search engine Google. The dates were saved in a special numerical chart of a data bank which is structured on the basis of word classes. The diploma thesis comes to the conclusion that each adjectival entry can be carried out in more ways depending on the preferred criterion which can be frequency based, syntactical, semantical or didactical. In the practical part of the...

Made available in DSpace on 2016-03-22T17:26:45Z (GMT). No. of bitstreams: 1 Magnun_DISSERTACAO_FINAL.pdf: 1540296 bytes, checksum: fc71cd85806540b80d03976dbc853af3 (MD5) Previous issue date: 2011-11-19 / The phenomenon of CV co-occurrence is characterized by the prevalent combination of some consonants and some vowels during the period of babbling along with the first words. Two main perspectives based on a dynamic origin have been evoked to explain the issue. The first one is called the Frame-then-Content theory (MACNEILAGE, 1998), which predicts that biomechanical inertia is the sensorimotor base of the syllable, the frame, preceding the phonic content. Under such view, the non-communicative movements would be the frame to the emergence of linguistic content phonic and the relationship between the movement and its acoustic effects would have privileged the speech, in the phylogenesis, as the main way to communication. The second perspective (ALBANO, 2009 a, b; 2011) states that the phonic factors in charge of forming the CV co-occurrence interact with individual differences and environmental conditions towards the child s linguistic production. In this work, we aim at (1) investigating the role played by the consonantal class of the sonorants and obstruents in the formation of CV coordination relations during the acquisition of BP. From this main goal, we have carried out a research on patterns of CV co-occurrence in longitudinal data of BP acquisition by 10 children, whose age was between 1:0 and 3:0 in the LIDES database, located in four age groups. We have also investigated the phenomenon in three stressing conditions unstressed initial syllable, stressed initial syllable and unstressed final syllable in an age group varying from 1:5 up to 2:5. Results show that child development seems to depend on the dynamics and motility in general. Such fact is evidenced by the changes regarding place of stressing and frequency of types and occurrences. This finding suggests that linguistic aspects, possibly determined be the grammar of the language, come into action more significantly in these positions, opposed to the claims of the Frame-and-Content theory. With respect to the role of the consonantal classes, we have found that the formation of the CV co-occurrences is strongly determined by the infant lexicon, which has been proved by the frequency of types and occurrences / O fenômeno da coocorrência CV caracteriza-se pela combinação prevalente de algumas consoantes com algumas vogais no período do balbucio e primeiras palavras. Duas vertentes principais, de origem dinâmica, têm sido utilizadas para explicar o fenômeno. A primeira delas é a teoria Frame-then-Content (MACNEILAGE, 1998), que prediz ser a inércia biomecânica a base sensório-motora da sílaba, o molde, precedente ao conteúdo fônico. Nessa linha, os movimentos não comunicativos seriam o molde para a emergência de conteúdo linguístico - fônico - e a relação do movimento com seus efeitos acústicos teria privilegiado a fala, na filogênese, como principal via de comunicação. A segunda vertente (ALBANO, 2009 a,b; 2011) afirma que os fatores fônicos formadores da coocorrência CV interagem com diferenças individuais e condições ambientais para a produção linguística da criança. Neste trabalho, objetivamos (1) investigar o papel desempenhado pela classe consonantais das soantes e das obstruintes na formação das relações de coordenação CV durante a aquisição do PB. A partir desse objetivo geral, pesquisamos os padrões de coocrrência CV em dados longitudinais de aquisição do PB de 10 crianças cujas faixas etárias situam-se entre 1:0 e 3:0 do Banco de dados LIDES, divididos em quatro grupos etários. Pesquisamos também fenômeno em três posições acentuais átona inicial, tônica e átona final em grupo etário de 1:5 a 2:5. Os resultados encontrados evidenciam que o desenvolvimento infantil parece ser subordinado a princípios da dinâmica e da motricidade em geral. Tal fato é evidenciado pela modificação dos vieses de combinações conforme a posição acentual investigada e pela frequência de tipos e ocorrências. Esse achado indica que aspectos linguísticos, determinados possivelmente pela gramática da língua, entram em ação mais fortemente nessas posições, fato que contraria as previsões da teoria Molde/Conteúdo. Em relação ao papel das classes consonantais, verificamos que a formação das coocorrências CV é determinada principalmente pelo léxico infantil, fato constatado pela frequência de tipos e ocorrências

Les écosystèmes littoraux lacustres, ou ceintures des lacs, ont une valeur écologique très importante. Zones de transition et d'échanges entre les écosystèmes terrestre et aquatique, les rives des lacs et leurs zones littorales sont reconnues pour abriter une forte biodiversité, dont des plantes aquatiques visibles à l'œil nu : les macrophytes. La structure spatiale, l'abondance et la composition spécifique de ces peuplements végétaux sont conditionnées par le caractère et le fonctionnement physiques des biotopes littoraux lacustres. Ce travail de thèse est dédié à l'acquisition de connaissances sur les relations entre les peuplements macrophytiques et les déterminants physiques de biotopes littoraux lacustres dans les lacs peu profonds. Il s'applique à définir, à différentes échelles, les préférences hydroécologiques et la structure fonctionnelle et spatiale de communautés et populations de plantes aquatiques. Cette recherche s'appuie sur la modélisation et sur des observations, mesures et prélèvements menés in situ sur les macrophytes et les variables environnementales dans les grands lacs naturels du littoral aquitain. Les principaux résultats indiquent que l'action du vent et des vagues, la pente des rives, la profondeur, la granulométrie et la matière organique des sédiments influencent l'organisation spatiale et les assemblages d'espèces de macrophytes. Ces variables peuvent être utilisées pour prédire la distribution potentielle de certaines communautés végétales. Les altérations anthropiques de l'hydromorphologie favorisent le développement des espèces exotiques tout en compromettant le maintien des espèces indigènes patrimoniales. Les enjeux liés à la gestion des espèces exotiques et des espèces indigènes ont également été identifiés à partir d'une analyse diachronique des dynamiques écologiques de la végétation depuis quarante ans dans les lacs aquitains. Il est désormais urgent d'établir une stratégie de conservation des espèces indigènes, voire de restauration pour les patrimoniales, et de limiter la dispersion des espèces exotiques dans les lacs. Présentant de nombreuses analogies avec les lacs d'Europe du Nord tels que les softwater lakes ou les Lobelia lakes, les processus écologiques en cours dans les lacs du sud-ouest de la France pourraient préfigurer la situation à venir pour l'ensemble des lacs du même type dans le contexte du changement global. Les écosystèmes littoraux lacustres concentrent des enjeux écologiques, économiques et sociétaux, soulevant des questions et problématiques scientifiques sur leur fonctionnement et sur la bioindication de leur état écologique qu'il est nécessaire de continuer à explorer. / Lakeshores zones are of very high ecological value. As they are ecotones between terrestrial and aquatic ecosystems, lake shoreline and littoral zones are characterized by a high level of biodiversity, including macroscopic aquatic plants known as macrophytes. Plant distribution, abundance and species composition are determined by physical features of lakeshores. This work aims to assess relationships between macrophyte populations and communities and hydromorphological components in shallow lakes at different scales. Plant and environmental variables were modeled or collected on the field in large lakes located along the western French Atlantic coast. Wind and wave action, slope, depth, sediment grain size and organic matter content are well known for their influence on macrophyte species distribution and assemblages. Physical variables can be used to predict the potential distribution of plant communities. Anthropogenic physical disturbances enhances the development of invasive species while they threaten rare and endangered native species. Plant management issues were also highlighted from a long-term field data (fourty years) analysis of macrophyte dynamics in these shallow lakes. Management strategies are needed for the conservation of native species, or even their restoration, and the control of alien species spread in these lakes. Similar in some ways to northern European lakes such as softwater lakes or Lobelia lakes, the ecological processes identified in these Atlantic shallow lakes could foreshadow the future situation for all lakes of the same type in the context of global change. Lakeshore ecosystems gather ecological, economic and societal issues, raising scientific questions and issues about their functioning and the bioindication of their ecological status that needs to be further explored.

This thesis looks at the evaluation of textures in two different perspectives using logarithmic image processing (LIP) framework. The first case after introducing the concept of textures and giving some classical approaches of textures evaluation, it gives an original approach of textures evaluation called covariogram which is derived from similarity metrics like distances or correlations etc. The classical covariogram which is derived from the classical similarity metrics and LIP covariogram are then applied over several images and the efficiency of the LIP one is clearly shown for darkened images. The last two chapters offer a new approach by considering the gray levels of an image as the phases of a medium. Each phase simulates like a percolation of a liquid in a medium defining the percolation trajectories. The propagation from one pixel to another is taken as easy or difficult determined by the difference of the gray level intensities. Finally different parameters like fractality from fractal dimensions, mean histogram etc associated to these trajectories are derived, based on which the primary experiment for the classification of random texture is carried out determining the relevance of this idea. Obviously, our study is only first approach and requires additional workout to obtain a reliable method of classification

This diploma thesis occupies itself with the subject of near synonymy, concretely with adjectives. On the basis of corpus linguistic methods two pairs of near synonyms have been researched - verschieden/unterschiedlich and bedeutend/bedeutsam. The 15 primary collocators (according to the syntactic position of each adjective) have been examined using the InterCorp parallel corpus methods in order to find out the most frequent Czech equivalence. Keywords: lexical-semantic relations, near synonymy, lexicography, corpora, cooccurrence analysis, Self Organizing Maps, CCDB

No description available.

The rate at which articles gets published grows exponentially and the possibility to access texts in machine-readable formats is also increasing. The need of an automated system to gather relevant information from text, text mining, is thus growing. The goal of this thesis is to find a biologically relevant gene network for atherosclerosis, themain cause of cardiovascular disease, by inspecting gene cooccurrences in abstracts from PubMed. In addition to this gene nets for yeast was generated to evaluate the validity of using text mining as a method. The nets found were validated in many ways, they were for example found to have the well known power law link distribution. They were also compared to other gene nets generated by other, often microbiological, methods from different sources. In addition to classic measurements of similarity like overlap, precision, recall and f-score a new way to measure similarity between nets are proposed and used. The method uses an urn approximation and measures the distance from comparing two unrelated nets in standard deviations. The validity of this approximation is supported both analytically and with simulations for both Erd¨os-R´enyi nets and nets having a power law link distribution. The new method explains that very poor overlap, precision, recall and f-score can still be very far from random and also how much overlap one could expect at random. The cutoff was also investigated. Results are typically in the order of only 1% overlap but with the remarkable distance of 100 standard deviations from what one could have expected at random. Of particular interest is that one can only expect an overlap of 2 edges with a variance of 2 when comparing two trees with the same set of nodes. The use of a cutoff at one for cooccurrence graphs is discussed and motivated by for example the observation that this eliminates about 60-70% of the false positives but only 20-30% of the overlapping edges. This thesis shows that text mining of PubMed can be used to generate a biologically relevant gene subnet of the human gene net. A reasonable extension of this work is to combine the nets with gene expression data to find a more reliable gene net.

Le système conceptuel humain est connu pour contenir deux types de concepts principaux: concrets et abstraits. Les concepts abstraits tels que l'opinion ou la détermination expriment les relations séquentielles entre les entités, ainsi que les états mentaux et introspectifs qui caractérisent la conscience humaine. Les recherches antérieures se sont très peu intéressées à la manière dont les concepts abstraits sont représentés tant sur le plan cognitif que conceptuel. Dans la littérature, la représentation, récupération et traitement des concepts abstraits dans le système conceptuel sont principalement attribués au phénomène connu sous le nom de l’effet de concrétude (avantage pour les mots concrets relativement aux mots abstraits par rapport aux processus cognitifs). Les théories actuelles de la cognition incarnée telles que la Théorie des Symboles Perceptifs proposent que les expériences réelles et les informations situationnelles pourraient jouer un rôle clé dans la façon dont les gens simulent, comprennent et utilisent les concepts abstraits. Le but principal de la présente thèse était d’explorer la structure conceptuelle, l'organisation et la représentation des concepts abstraits dans le système cognitif. Pour ce faire, quatre séries d'études expérimentales utilisant des tâches de catégorisation et de jugement de similarité ont été réalisées. Le premier objectif était de déterminer l’effet des informations situationnelles sur la réorganisation des concepts abstraits. Le deuxième objectif était de déterminer si une organisation taxonomique ou thématique pourrait être à la base de la représentation conceptuelle des concepts abstraits. Les résultats suggèrent que la réorganisation thématique et les informations situationnelles jouent un rôle central dans le traitement et la réorganisation conceptuels des concepts abstraits. / The human conceptual system is known to contain two main types of concepts: concrete and abstract. Abstract concepts such as opinion or determination express the sequences of relations between different entities. They also manifest the internal and introspective states of existence that characterize the human consciousness. The semantic representation and organization of abstract concepts has received very little attention in the cognitive psychology literature over the past decades, whereas the vast majority of studies have been dedicated to concrete concepts. Previous research on abstract concepts has explained how they are conceptually represented by focusing on their differences from concrete concepts i.e., the concreteness effect. Current theories of grounded cognition such as the Perceptual Symbol Systems Theory propose that situational knowledge and experiences could play a key role in how people simulate, understand and use abstract concepts.Our aim was to assess the principles that underlie the conceptual structure, organization and representation of abstract concepts within the cognitive system. Four series of behavioural experiments using categorization and similarity judgment tasks were designed to investigate the role of situational information and thematic organization in the processing of abstract concepts. The results indicated that the co-occurrences and the experiencing of unrelated abstract concepts in relevant situations significantly influenced the emergence of novel thematic reorganizations between the concepts compared to baseline. Thus, suggesting the central role that thematic reorganization and situational information play in the conceptual representation of abstract concepts.

En fait, le concept de texture n’est pas facile à définir, mais il est clair qu’il est fortement lié au Système Visuel Humain. Sachant que le Modèle LIP est compatible avec la vision humaine, il nous a semblé intéressant de créer des outils logarithmiques dédiés à l’évaluation de la texture. Nous nous sommes concentrés sur la notion de covariogramme, qui peut être pilotée par diverses métriques logarithmiques. Ces métriques jouent le rôle d’outils de “corrélation”, avec l’avantage de prendre en compte la vision humaine. De plus, les outils LIP sont peu dépendants des conditions d’éclairement et fournissent donc des résultats robustes si celles-ci varient. Les deux derniers Chapitres proposent une nouvelle approche consistant à considérer les niveaux de gris d’une image comme les phases d’un milieu. Chaque phase permet de simuler la percolation d’un liquide dans le milieu, définissant ainsi des trajectoires de percolation. Chaque propagation d’un pixel à un autre est considérée comme facile ou non, en fonction des niveaux de gris traversés. Une « fonction de coût » est créée, qui modifie le « temps » de propagation d’un point à l’autre. De plus, la fonction de coût peut être calculée dans le contexte LIP, pour prendre en compte la vision humaine / This thesis looks at the evaluation of textures in two different perspectives using logarithmic image processing (LIP) framework. The first case after introducing the concept of textures and giving some classical approaches of textures evaluation, it gives an original approach of textures evaluation called covariogram which is derived from similarity metrics like distances or correlations etc. The classical covariogram which is derived from the classical similarity metrics and LIP covariogram are then applied over several images and the efficiency of the LIP one is clearly shown for darkened images. The last two chapters offer a new approach by considering the gray levels of an image as the phases of a medium. Each phase simulates like a percolation of a liquid in a medium defining the percolation trajectories. The propagation from one pixel to another is taken as easy or difficult determined by the difference of the gray level intensities. Finally different parameters like fractality from fractal dimensions, mean histogram etc associated to these trajectories are derived, based on which the primary experiment for the classification of random texture is carried out determining the relevance of this idea. Obviously, our study is only first approach and requires additional workout to obtain a reliable method of classification

Cette thèse traite des problèmes de modélisation, reconnaissance, localisation et utilisation du contexte pour la manipulation d'objets par un robot. Le processus de modélisation se divise en quatre composantes : le système réel, les données capteurs, les propriétés à reproduire et le modèle. En spécifiant chacune des ces composantes, il est possible de définir un processus de modélisation adapté au problème présent, la manipulation d'objets par un robot. Cette analyse mène à l'adoption des descripteurs de texture locaux pour la modélisation. La modélisation basée sur des descripteurs de texture locaux a été abordé dans de nombreux travaux traitant de structure par le mouvement (SfM) ou de cartographie et localisation simultanée (SLAM). Les méthodes existantes incluent Bundler, Roboearth et 123DCatch. Pourtant, aucune de ces méthodes n'a recueilli le consensus. En effet, l'implémentation d'une approche similaire montre que ces outils sont difficiles d'utilisation même pour des utilisateurs experts et qu'ils produisent des modèles d'une haute complexité. Cette complexité est utile pour fournir un modèle robuste aux variations de point de vue. Il existe deux façons pour un modèle d'être robuste : avec le paradigme des vues multiple ou celui des descripteurs forts. Dans le paradigme des vues multiples, le modèle est construit à partir d'un grand nombre de points de vue de l'objet. Le paradigme des descripteurs forts compte sur des descripteurs résistants aux changements de points de vue. Les expériences réalisées montrent que des descripteurs forts permettent d'utiliser un faible nombre de vues, ce qui résulte en un modèle simple. Ces modèles simples n'incluent pas tout les point de vus existants mais les angles morts peuvent être compensés par le fait que le robot est mobile et peut adopter plusieurs points de vue. En se basant sur des modèles simples, il est possible de définir des méthodes de modélisation basées sur des images seules, qui peuvent être récupérées depuis Internet. A titre d'illustration, à partir d'un nom de produit, il est possible de récupérer des manières totalement automatiques des images depuis des magasins en ligne et de modéliser puis localiser les objets désirés. Même avec une modélisation plus simple, dans des cas réel ou de nombreux objets doivent être pris en compte, il se pose des problèmes de stockage et traitement d'une telle masse de données. Cela se décompose en un problème de complexité, il faut traiter de nombreux modèles rapidement, et un problème d'ambiguïté, des modèles peuvent se ressembler. L'impact de ces deux problèmes peut être réduit en utilisant l'information contextuelle. Le contexte est toute information non issue des l'objet lui même et qui aide a la reconnaissance. Ici deux types de contexte sont abordés : le lieu et les objets environnants. Certains objets se trouvent dans certains endroits particuliers. En connaissant ces liens lieu/objet, il est possible de réduire la liste des objets candidats pouvant apparaître dans un lieu donné. Par ailleurs l'apprentissage du lien lieu/objet peut être fait automatiquement par un robot en modélisant puis explorant un environnement. L'information appris peut alors être fusionnée avec l'information visuelle courante pour améliorer la reconnaissance. Dans les cas des objets environnants, un objet peut souvent apparaître au cotés d'autres objets, par exemple une souris et un clavier. En connaissant la fréquence d'apparition d'un objet avec d'autres objets, il est possible de réduire la liste des candidats lors de la reconnaissance. L'utilisation d'un Réseau de Markov Logique est particulièrement adaptée à la fusion de ce type de données. Cette thèse montre la synergie de la robotique et du contexte pour la modélisation, reconnaissance et localisation d'objets. / This Thesis addresses the modeling, recognition, localization and use of context for objects manipulation by a robot. We start by presenting the modeling process and its components: the real system, the sensors' data, the properties to reproduce and the model. We show how, by specifying each of them, one can define a modeling process adapted to the problem at hand, namely object manipulation by a robot. This analysis leads us to the adoption of local textured descriptors for object modeling. Modeling with local textured descriptors is not a new concept, it is the subject of many Structure from Motion (SfM) or Simultaneous Localization and Mapping (SLAM) works. Existing methods include bundler, roboearth modeler and 123DCatch. Still, no method has gained widespread adoption. By implementing a similar approach, we show that they are hard to use even for expert users and produce highly complex models. Such complex techniques are necessary to guaranty the robustness of the model to view point change. There are two ways to handle the problem: the multiple views paradigm and the robust features paradigm. The multiple views paradigm advocate in favor of using a large number of views of the object. The robust feature paradigm relies on robust features able to resist large view point changes. We present a set of experiments to provide an insight into the right balance between both. By varying the number of views and using different features we show that small and fast models can provide robustness to view point changes up to bounded blind spots which can be handled by robotic means. We propose four different methods to build simple models from images only, with as little a priori information as possible. The first one applies to planar or piecewise planar objects and relies on homographies for localization. The second approach is applicable to objects with simple geometry, such as cylinders or spheres, but requires many measures on the object. The third method requires the use of a calibrated 3D sensor but no additional information. The fourth technique doesn't need a priori information at all. We apply this last method to autonomous grocery objects modeling. From images automatically retrieved from a grocery store website, we build a model which allows recognition and localization for tracking. Even using light models, real situations ask for numerous object models to be stored and processed. This poses the problems of complexity, processing multiple models quickly, and ambiguity, distinguishing similar objects. We propose to solve both problems by using contextual information. Contextual information is any information helping the recognition which is not directly provided by sensors. We focus on two contextual cues: the place and the surrounding objects. Some objects are mainly found in some particular places. By knowing the current place, one can restrict the number of possible identities for a given object. We propose a method to autonomously explore a previously labeled environment and establish a correspondence between objects and places. Then this information can be used in a cascade combining simple visual descriptors and context. This experiment shows that, for some objects, recognition can be achieved with as few as two simple features and the location as context. The objects surrounding a given object can also be used as context. Objects like a keyboard, a mouse and a monitor are often close together. We use qualitative spatial descriptors to describe the position of objects with respect to their neighbors. Using a Markov Logic Network, we learn patterns in objects disposition. This information can then be used to recognize an object when surrounding objects are already identified. This Thesis stresses the good match between robotics, context and objects recognition.