Dans cette thèse nous avons exploré la capacité des modèles magnétiques de la physique statistique à extraire l'information essentielle contenue dans les textes. Les documents ont été représentés comme des ensembles d'unités en interaction magnétique, l'intensité de telles interactions a été mesurée et utilisée pour calculer de quantités qui sont des indices de l'importance de l'information portée. Nous proposons deux nouvelles méthodes. Premièrement, nous avons étudié un modèle de spins qui nous a permis d'introduire l'énergie textuelle d'un document. Cette quantité a été utilisée comme indicatrice de pertinence et appliquée à une vaste palette de tâches telles que le résumé automatique, la recherche d'information, la classification de documents et la segmentation thématique. Par ailleurs, et de façon encore exploratoire, nous proposons un deuxième algorithme qui définie un couplage grammatical pour conserver les termes importants et produire des contractions. De cette façon, la compression d'une phrase est l'état fondamental de la chaîne de termes. Comme cette compression n'est pas forcement bonne, il a été intéressant de produire des variantes en permettant des fluctuations thermiques. Nous avons fait des simulations Métropolis Monte-Carlo avec le but de trouver l'état fondamental de ce système qui est analogue au verre de spin. / In this thesis we explored the ability of magnetic models of statistical physics to extract the essential information contained in texts. Documents are represented as sets of interacting magnetic units, the intensity of such interactions are measured and they are used to calculate quantities that are evidence of the importance of information scope. We propose two new methods. Firstly, we studied a spin model which allowed us to introduce the textual energy. This quantity was used as an indicator of information relevance. Several adaptations were necessary to adapt the energy calculation to a wide range of tasks such as summarisation, information retrieval, document classification and thematic segmentation. Furthermore, and even exploratory, we propose a second algorithm that defines a grammatical coupling between types of terms to retain the important terms and produce contractions. In this way, the compression of a sentence is the ground state of the chain of terms. As this compression is not necessarily good, it was interesting produce variants by thermal fluctuations. We have done simulations Metropolis Monte-Carlo with the aim of finding the ground state of this system that is analogous to spin glass.

No description available.

No description available.

La recherche d'images basée sur le contenu visuel est un domaine très actif de la vision par ordinateur, car le nombre de bases d'images disponibles ne cesse d'augmenter.L’objectif de ce type d’approche est de retourner les images les plus proches d'une requête donnée en terme de contenu visuel.Notre travail s'inscrit dans un contexte applicatif spécifique qui consiste à indexer des petites bases d'images expertes sur lesquelles nous n'avons aucune connaissance a priori.L’une de nos contributions pour palier ce problème consiste à choisir un ensemble de descripteurs visuels et de les placer en compétition directe. Nous utilisons deux stratégies pour combiner ces caractéristiques : la première, est pyschovisuelle, et la seconde, est statistique.Dans ce contexte, nous proposons une approche adaptative non supervisée, basée sur les sacs de mots et phrases visuels, dont le principe est de sélectionner les caractéristiques pertinentes pour chaque point d'intérêt dans le but de renforcer la représentation de l'image.Les tests effectués montrent l'intérêt d'utiliser ce type de méthodes malgré la domination des méthodes basées réseaux de neurones convolutifs dans la littérature.Nous proposons également une étude, ainsi que les résultats de nos premiers tests concernant le renforcement de la recherche en utilisant des méthodes semi-interactives basées sur l’expertise de l'utilisateur. / Image Retrieval is still a very active field of image processing as the number of available image datasets continuously increases.One of the principal objectives of Content-Based Image Retrieval (CBIR) is to return the most similar images to a given query with respect to their visual content.Our work fits in a very specific application context: indexing small expert image datasets, with no prior knowledge on the images. Because of the image complexity, one of our contributions is the choice of effective descriptors from literature placed in direct competition.Two strategies are used to combine features: a psycho-visual one and a statistical one.In this context, we propose an unsupervised and adaptive framework based on the well-known bags of visual words and phrases models that select relevant visual descriptors for each keypoint to construct a more discriminative image representation.Experiments show the interest of using this this type of methodologies during a time when convolutional neural networks are ubiquitous.We also propose a study about semi interactive retrieval to improve the accuracy of CBIR systems by using the knowledge of the expert users.

Cette thèse, consacrée à l’expression de l’interdiction en russe moderne, étudie les différents types d’énoncés prohibitifs. L’acte de langage d’interdiction correspond à un énoncé performatif et contraignant incitant au non-accomplissement d’une action perçue comme indésirable par le locuteur. Cet énoncé crée, pour l’allocutaire, une obligation qui jusque-là n’existait pas. L’expression de l’interdiction se fait à différents niveaux de communication et par des moyens linguistiques de natures diverses : pragmatique, syntaxique ou lexicale. Outre les énoncés performatifs, l’interdiction peut être exprimée par des énoncés comportant une valeur illocutoire d’interdiction, y compris par des actes de langage indirects.Si les types d’énoncés prohibitifs sont nombreux, ils ne sont pas interchangeables. D’une part, ils ne peuvent être employés indifféremment dans n’importe quel contexte et ils relèvent de niveaux de langues différents. D’autre part, les disparités existant dans la force illocutoire des différents énoncés prohibitifs ont été mises en évidence. La force illocutoire d’un énoncé prohibitif résulte du choix du type d’énoncé retenu, mais aussi des moyens linguistiques qui peuvent être employés pour le moduler.Cette étude a permis d’établir une typologie des énoncés prohibitifs. Ces énoncés peuvent se répartir en trois catégories : les énoncés qui signifient l’interdiction par exception, les énoncés habituellement prohibitifs et les énoncés invariablement prohibitifs.S’y ajoute la question de l’évolution de la langue, notamment par un effet de contamination des langues européennes. Cette évolution se traduit par le recours moins fréquent aux tournures les plus catégoriques, ainsi que par l’introduction de formules de politesse de plus en plus fréquentes. / This thesis is dedicated to prohibitive commands in modern Russian and studies the various speech strategies expressing negative commands. Prohibitive speech acts correspond to compulsory and performative utterances urging the non-accomplishment of an action perceived as undesirable by the speaker. Due to this utterance, the addressee has no other option than compliance, which was not the case before.Prohibitive commands are expressed at various communication levels and by a wide range of linguistic means: pragmatic, syntactic or lexical.Besides performative utterances, prohibition can be expressed by utterances with a prohibitive illocutionary force, including indirect speech acts. Although negative commands can be expressed by various strategies, they are not interchangeable. On the one hand, any negative command cannot be used in all contexts and in addition, the registers are different.On the other hand, the disparities in the illocutionary force of command of the various prohibitive strategies could be established. The illocutionary force of a prohibitive command both results from the selected linguistic strategy, but also from the means of modulation that can be used.This study enabled us to establish a typology of negative commands. The latter can be displayed in three categories: utterances whose prohibitive meaning occurs exceptionally, usually prohibitive utterances and invariably prohibitive utterances. Furthermore, the evolution of the Russian language has to be taken into account, among others by a contamination effect from European languages. This evolution leads to a less frequent resort to the strongest prohibitive commands, as well as to the increasing use of politeness strategies.

Lai Wing-sze. / Thesis submitted in: August 2004. / Thesis (M.Phil.)--Chinese University of Hong Kong, 2005. / Includes bibliographical references (leaves 233-237). / Abstracts in English and Chinese. / Abstract --- p.i / Acknowledgements --- p.iii / Chapter Chapter 1 --- Introduction --- p.1 / Chapter 1.1 --- Rational and Significance --- p.1 / Chapter 1.2 --- Scope of analysis --- p.2 / Chapter 1.3 --- Organization --- p.4 / Chapter Chapter 2 --- Background --- p.6 / Chapter 2.1 --- Comparison of Cantonese and English consonants --- p.6 / Chapter 2.2 --- Comparison of Cantonese and English vowels --- p.8 / Chapter 2.3 --- Comparison of Cantonese and English syllables --- p.9 / Chapter 2.4 --- Comparison of Cantonese and English autosegments --- p.12 / Chapter Chapter 3 --- Literature Review --- p.18 / Chapter 3.1 --- Tone assignment in Cantonese loanwords --- p.18 / Chapter 3.2 --- Relationship between tone and stress --- p.27 / Chapter 3.3 --- Research Questions --- p.31 / Chapter Chapter 4 --- Methodology --- p.32 / Chapter 4.1 --- Data Collection and Classification --- p.32 / Chapter 4.2 --- Loanword Verification Task --- p.48 / Chapter 4.3 --- Loanword Creation Task --- p.51 / Chapter Chapter 5 --- Investigation of Tonal Patterns --- p.55 / Chapter 5.1 --- Monosyllabic Loanwords --- p.55 / Chapter 5.2 --- Disyllabic Loanwords --- p.59 / Chapter 5.3 --- Summary of Tonal Patterns in Monosyllabic and Disyllabic Loanwords --- p.81 / Chapter 5.4 --- Loanwords with more than two syllables --- p.81 / Chapter Chapter 6 --- Further Evidence on Tonal Patterns --- p.92 / Chapter 6.1 --- Results of Loanword Verification Task --- p.92 / Chapter 6.2 --- Results of Loanword Creation Task --- p.103 / Chapter 6.3 --- Tonal Assignment by New Generation --- p.115 / Chapter Chapter 7 --- Constraints and Rankings in Tone Assignment in Loanwords --- p.116 / Chapter 7.1 --- OT Concepts and Mechanisms --- p.116 / Chapter 7.2 --- Tonal Features --- p.120 / Chapter 7.3 --- My Proposed Account --- p.129 / Chapter 7.4 --- Constraints and Rankings --- p.132 / Chapter 7.5 --- Summary --- p.145 / Chapter Chapter 8 --- Conclusion and Implication --- p.147 / Appendices --- p.150 / Appendix I All collected Loanword Data --- p.150 / Appendix II Eliminated Loanword Data --- p.164 / Appendix III Kept Loanword Data --- p.168 / Appendix IV Initial Loanword Classification --- p.181 / Appendix V Loanword Questionnaire A --- p.198 / Appendix VI Loanword Questionnaire A Script --- p.201 / Appendix VII Loanword Questionnaire B --- p.203 / Appendix VIII More Examples of σ→ σ55 --- p.207 / Appendix IX More Examples of 'σσ --- p.209 / Appendix X More Examples of'σ[ --- p.214 / Appendix XI More Examples of ['σ --- p.217 / Appendix XII Disyllabic Loanwords with Atypical Tonal Patterns --- p.218 / Appendix XIII Disyllabic Loanwords with Other Tonal Influences --- p.220 / Appendix XIV Trisyllabic Loanwords with Atypical Tonal Patterns --- p.222 / Appendix XV Results of Loanword Creation Task --- p.224 / References --- p.233

No description available.

This dissertation analyses when, how and why Turkish loanwords became incorporated into Bosnian and Bulgarian, as seen in the writings of the seventeenth- and eighteenth-century Bosnian and Bulgarian Franciscans. I analyse Bosnian works (religious and secular) by Matija Divkovic, Ivan Bandulavic, Pavo Posilovic Mošunjanin, Mihovil Radnic, Stjepan Margitic Markovac, Lovro Braculjevic, Filip Lastric, Nikola Marcinkušic Lašvanin, Marko Dobretic, Bono Benic, and Grgo Ilijic-Varešanin. As a Bulgarian counterpart, I analyse three eighteenth-century Bulgarian Franciscan manuscripts and the works of Petar Bogdan Bakšic and Filip Stanislavov. The dissertation consists of eight chapters. The first chapter gives background information on Turkish presence in Bosnia and Bulgaria, the history of the Franciscans in Bosnia and Bulgaria, short biographies of each of the writers whose works are analysed, phonology and orthography. The second chapter focuses on the complications regarding establishing earliest attestations for turkisms in Bosnian and Bulgarian. The third chapter discusses the nominal morphology of turkisms in Bosnian and Bulgarian. This chapter analyses why turkisms developed the gender that they did when borrowed from a language that does not have gender as a category. Chapter four addresses the verbal morphology of turkisms in Bosnian and Bulgarian. Verbal prefixes are discussed in detail, as are Turkish voiced suffixes in Bulgarian. The fifth chapter analyses adjectives and adverbs, with focus on gender and number agreement. The sixth chapter addresses the use of Turkish conjunctions. The seventh chapter looks at the motivation, semantics and setting of turkisms in Bosnian and Bulgarian. The conclusion addresses how morphology, semantics, motivation and setting of turkisms relate to their chronology in Bosnian and Bulgarian and how these areas differ from language to language.

The object of this dissertation is the expression of space in Mikalojus Daukša’s Postilla (1599). It is limited to the semantic analysis of the four postpositional local cases, their distribution, and their relationship to their functional equivalents (prepositions, genitive and dative cases). The research is based on the conception of linguistic units as natural categories. Postpositional locatives are treated as multifunctional cases possessing a number of various use types that differ in status: some are prototypical, other are located in the periphery of the case. A semasiological analysis of each postpositional local case is performed starting from their concrete spatial use types and then going into their abstract domains. The use types and category structure of each postpositional locative are systematically described. The analysis of each case is followed by an onomasiological investigation aiming to analyze the semantic distribution between prepositional phrases and postpositional locatives. It is shown that they profile different aspects of the encoded spatial scenes. Since Daukša’s Postilla is a translation from Polish (Postilla katolicka mnieysza by Jakob Wujek, 1590), the influence of the Polish original is also evaluated. / Disertacijoje aprašoma vietos raiška Mikalojaus Daukšos Postilėje (1599), dėl plačios vietos raiškos priemonių sistemos apsiribojant postpozicinių vietininkų ir jų funkcinių ekvivalentų (prielinksninių junginių ar vietos reikšme vartojamų linksnių) analize. Tyrime laikomasi kognityvinės poliseminių kalbos vienetų reikšmės sampratos: vietininko reikšmę sudaro tarpusavyje susiję vartosenos tipai, kurių vieni yra būdingesni, prototipiškesni, o kiti – esantys linksnio reikšmių periferijoje. Disertacijoje sistemiškai aprašomos visos postpozicinių vietininkų reikšmės ir išskiriami vietos reikšmės tipai: tai leidžia verifikuoti tradicinį požiūrį, kad postpoziciniai vietininkai yra vietos linksniai. Atlikus kiekvieno postpozicinio vietininko semasiologinę analizę (išskyrus vartosenos tipus bei nustačius, kiek jų susiję su vietos raiška), einama onomasiologine kryptimi – apibrėžiamas kiekvieno postpozicinio vietininko ir jo funkcinių ekvivalentų santykis, išryškinant kalbamųjų raiškos priemonių semantinius skirtumus. Disertacijoje nustatyta vertimo šaltinio (Jokūbo Wujeko Postilla katolicka mnieysza, 1590) įtaka Daukšos Postilės vietos raiškos sintagmoms, apibrėžtas postpozicinių vietininkų tarpusavio santykis ir įvertinti literatūroje nurodomi šių linksnių painiojimo atvejai.

Disertacijoje aprašoma vietos raiška Mikalojaus Daukšos Postilėje (1599), dėl plačios vietos raiškos priemonių sistemos apsiribojant postpozicinių vietininkų ir jų funkcinių ekvivalentų (prielinksninių junginių ar vietos reikšme vartojamų linksnių) analize. Tyrime laikomasi kognityvinės poliseminių kalbos vienetų reikšmės sampratos: vietininko reikšmę sudaro tarpusavyje susiję vartosenos tipai, kurių vieni yra būdingesni, prototipiškesni, o kiti – esantys linksnio reikšmių periferijoje. Disertacijoje sistemiškai aprašomos visos postpozicinių vietininkų reikšmės ir išskiriami vietos reikšmės tipai: tai leidžia verifikuoti tradicinį požiūrį, kad postpoziciniai vietininkai yra vietos linksniai. Atlikus kiekvieno postpozicinio vietininko semasiologinę analizę (išskyrus vartosenos tipus bei nustačius, kiek jų susiję su vietos raiška), einama onomasiologine kryptimi – apibrėžiamas kiekvieno postpozicinio vietininko ir jo funkcinių ekvivalentų santykis, išryškinant kalbamųjų raiškos priemonių semantinius skirtumus. Disertacijoje nustatyta vertimo šaltinio (Jokūbo Wujeko Postilla katolicka mnieysza, 1590) įtaka Daukšos Postilės vietos raiškos sintagmoms, apibrėžtas postpozicinių vietininkų tarpusavio santykis ir įvertinti literatūroje nurodomi šių linksnių painiojimo atvejai. / The object of this dissertation is the expression of space in Mikalojus Daukša’s Postilla (1599). It is limited to the semantic analysis of the four postpositional local cases, their distribution, and their relationship to their functional equivalents (prepositions, genitive and dative cases). The research is based on the conception of linguistic units as natural categories. Postpositional locatives are treated as multifunctional cases possessing a number of various use types that differ in status: some are prototypical, other are located in the periphery of the case. A semasiological analysis of each postpositional local case is performed starting from their concrete spatial use types and then going into their abstract domains. The use types and category structure of each postpositional locative are systematically described. The analysis of each case is followed by an onomasiological investigation aiming to analyze the semantic distribution between prepositional phrases and postpositional locatives. It is shown that they profile different aspects of the encoded spatial scenes. Since Daukša’s Postilla is a translation from Polish (Postilla katolicka mnieysza by Jakob Wujek, 1590), the influence of the Polish original is also evaluated.