Rozprawy doktorskie na temat „Bash Translation”

Machine learning and data mining have provided plenty of tools for extracting knowledge from data. Yet, such knowledge may not be directly applicable to target applications and might need further manipulation: The knowledge might contain too much noise, or the target application may use a different representation or terminology. In this dissertation, we study three problems related to knowledge management and manipulation. First, given a knowledge base (KB) automatically extracted from the text, we explore how to refine it based on the dependencies among the possible KB instances and their confidence values. Second, when the target application to which we want to apply our knowledge uses a different schema, we explore how to translate the knowledge based on the mapping between the schemas. Sometimes, the mapping between two schemas can be discovered automatically, so the third problem we consider is whether we can find the mapping more accurately using the corresponding knowledge contained in the two schemas. We notice that a large fraction of data and knowledge can be represented in relational models, which can be formalized with first-order logic. Moreover, uncertainty is a common feature existing in these problems, e.g., the confidence values associated with the KB instances, the probabilistic knowledge rules to be translated, or the schemas not perfectly aligned with each other. Therefore, we adopt statistical relational learning, which combines first-order logic with probabilistic models, to resolve these problems. In particular, we use Markov logic networks (MLNs), which consist of sets of weighted first-order formulas. MLNs are a powerful and flexible language for representing hard and soft constraints of relational domains. We develop the MLN formulations for each of these problems, and we use the representation, inference and learning approaches in the literature with certain adaptations to solve them. The experiment results show that MLNs successfully provide solutions to these problems or achieve better performances than the existing methods. This dissertation includes previously published and unpublished coauthored material.

Proefschrift ter verkrijging van de graad van Doctor--Faculteit der godgeleerdheid--Leiden--Rijksuniversiteit, 1994.
Contient la translittération en caractères latins du texte de la traduction turque des livres de Samuel, et en annexe le fac-sim. du manuscrit, en caractères arabes, de cette trad. Remerciements, résumé et curriculum vitae en néerlandais. Bibliogr. p. 401-404.

Understanding translation efficiency is crucial to pharmaceutical companies that have invested substantial time and effort in engineering bacteria to produce recombinant proteins. While translation initiation and elongation have been studied intensively, much remains obscure in the subprocess of translation termination. We aim to understand how stop codons and the first 3’ flanking (+4) base affect translation termination efficiency. In chapter two, we hypothesized that stop codon usage of UAG and UGA is dependent on the abundance of their respective decoders, RF1 and RF2. We predicted and observed that bacterial species with high relative proportions of RF1 uses UAG more, and vice versa for UGA. In addition, the usage of UGA, not UAG, is always avoided in highly expressed genes. Thus, we argued against the claim made by a recent study that UAG is a minor stop codon in bacteria. The claim is incorrect because UAG does not meet the two criteria of a minor codon: i) it is most avoided in highly expressed genes, and ii) it corresponds to the least abundant decoder. Interestingly, we found that the proportion of RF2 decreases rapidly towards zero in species with high AT contents; this explains why UGA is reassigned to a sense codon in bacterial lineages with high AT content. In chapter three, we examined the role of the first downstream (+4) base Uracil in bacterial translation termination. The +4U is associated with a decrease in stop codon read-through in bacteria and yeast. We hypothesized that i) +4U enhances the termination efficiency of stop signals, and ii) +4U may serve to prevent stop codon misreading by near cognate tRNAs (nc_tRNAs). We predicted that i) +4U is preferred in highly expressed genes (HEGs) than lowly expressed genes (LEGs), and ii) +4U usage increases with the frequency of stop codon nc_tRNAs. We found +4U consistently over-represented in HEGs in contrast to LEGs; however, +4U usage in HEGs decreases in GC-rich species where most stop codons are UGA and UAG. In addition, +4U usage increases significantly with UAA usage in the known highly expressed ribosomal protein genes. These results suggest that +4U is a strong stop signal enhancer for UAA, not UAG or UGA. Furthermore, in HEGs, +4U usage also increases significantly with the abundance of UAA nc_tRNAs, suggesting that +4U increases UAA termination efficiency presumably by reducing misreading of UAA by nc_tRNAs.

Bitext alignment is the task of aligning a text in a source language and its translation in the target language. Aligning amounts to finding the translational correspondences between textual units at different levels of granularity. Many practical natural language processing applications rely on bitext alignments to access the rich linguistic knowledge present in a bitext. While the most predominant application for bitexts is statistical machine translation, they are also used in multilingual (and monolingual) lexicography, word sense disambiguation, terminology extraction, computer-aided language learning andtranslation studies, to name a few.Bitext alignment is an arduous task because meaning is not expressed seemingly across languages. It varies along linguistic properties and cultural backgrounds of different languages, and also depends on the translation strategy that have been used to produce the bitext.Current practices in bitext alignment model the alignment as a hidden variable in the translation process. In order to reduce the complexity of the task, such approaches suppose that a word in the source sentence is aligned to one word at most in the target sentence.However, this over-simplistic assumption results in asymmetric, one-to-many alignments, whereas alignments are typically symmetric and many-to-many.To achieve symmetry, two one-to-many alignments in opposite translation directions are built and combined using a heuristic.In order to use these word alignments in phrase-based translation systems which use phrases instead of words, a heuristic is used to extract phrase pairs that are consistent with the word alignment.In this dissertation we address both the problems of word alignment and phrase pairs extraction.We improve the state of the art in several ways using discriminative learning techniques.We present a maximum entropy (MaxEnt) framework for word alignment.In this framework, links are predicted independently from one another using a MaxEnt classifier.The interaction between alignment decisions is approximated using stackingtechniques, which allows us to account for a part of the structural dependencies without increasing the complexity. This formulation can be seen as an alignment combination method,in which the union of several input alignments is used to guide the output alignment. Additionally, input alignments are used to compute a rich set of feature functions.Our MaxEnt aligner obtains state of the art results in terms of alignment quality as measured by thealignment error rate, and translation quality as measured by BLEU on large-scale Arabic-English NIST'09 systems.We also present a translation quality informed procedure for both extraction and evaluation of phrase pairs. We reformulate the problem in the supervised framework in which we decide for each phrase pair whether we keep it or not in the translation model. This offers a principled way to combine several features to make the procedure more robust to alignment difficulties. We use a simple and effective method, based on oracle decoding,to annotate phrase pairs that are useful for translation. Using machine learning techniques based on positive examples only,these annotations can be used to learn phrase alignment decisions. Using this approach we obtain improvements in BLEU scores for recall-oriented translation models, which are suitable for small training corpora.

In this study I examine the genesis and compositional purpose of Mendelssohn's seven chorale cantatas which include "Christe, du Lamm Gottes" (1827); "Jesu, meine Freude" (1828); "Wer nur den lieben Gott läßt walten" (1829); "O Haupt voll Blut und Wunden" (1830); "Vom Himmel hoch" (1831); "Wir Glauben all an einem Gott" (1831); and "Ach Gott, vom Himmel sieh’ darein" (1832). I also re-examine the widely held belief that Mendelssohn's chorale cantatas were juvenilia, functioning only as compositional exercises to better learn the style of J.S. Bach. My examination includes a stylistic analysis of each cantata that shows a direct deviation from an imitation of Bach's musical syntax and form to a re-creation of the chorale cantata genre using nineteenth-century idioms. My approach to analysis is modeled after that of James Garratt, who has done a great deal of research on Mendelssohn's psalm settings. These works, like the chorale cantatas, have also been criticized for being compositional exercises that imitate Palestrina's musical syntax and form. Garratt believes that by using a literary method called translation theory, one can gain an unbiased assessment and evaluation of compositions that are based on pre-existent models. I have applied the tenets of translation theory as the foundation for my analysis to determine if the chorale cantatas were imitations of Bach's model or re-creations of the genre. This approach provides, as Garratt states, "a means of differentiating between the degree of affinity that links the works of Mendelssohn...and [his] models." The stylistic analysis also illustrates how the chorale cantatas are exemplars of Mendelssohn's compositional style. Unique characteristics such as Mendelssohn's frank religious convictions; his perceptive and creative treatment of chorales; his understanding of Baroque forms and styles; and his remarkable contrapuntal prowess are all beautifully exemplified in the chorale cantatas. I conclude that instead of being regarded as archaic, compositional exercises, Mendelssohn's chorale cantatas can and should be viewed as seminal works in his sacred choral-orchestral output.

In the last two decades, the volume and variety of biomedical data has dramatically increased. The data is heterogeneous and scattered across many resources. This produces bottlenecks in the analysis and extraction of knowledge from this sea of information. To overcome this hurdle, better catalogs that integrate different data types, offer easy access to users, and support automatic workflows, are needed. With this in mind, we have developed DisGeNET, a discovery platform that contains information on more than 17,000 genes related to over 14,000 diseases. We have used DisGeNET to study the properties of disease genes in the context of protein interaction networks. To produce an accurate analysis of the mesoscale properties of the human interactome, we first compared the network partitions generated by two popular clustering algorithms, to assess how this would impact the follow-up biological analysis. Using the best performing algorithm we then explored the network properties of disease genes. Then we evaluated the relationship between the network properties of different groups of disease genes and their tolerance to likely deleterious germline variants across human populations. Finally, we have developed a new network medicine approach to study disease comorbidities, and applied it to the analysis of COPD comorbidities.
Los avances tecnológicos de las últimas dos décadas han producido un incremento dramático en la cantidad y la diversidad de datos biomédicos disponibles. Este proceso ha ocurrido de manera fragmentada, y en consecuencia los datos se encuentran almacenados en distintos repositorios, lo cual impone barreras a la hora de integrarlos, analizarlos y extraer conocimiento a partir de ellos. Para superar estas barreras, es necesario contar con recursos computacionales que integren esta información, y ofrezcan un fácil acceso a la misma, permitiendo al mismo tiempo su análisis automatizado. En respuesta a esta necesidad hemos desarrollado DisGeNET, una plataforma orientada a la exploración de las causas genéticas de las enfermedades humanas, que contiene actualmente información sobre más de 14.000 enfermedades y 17.000 genes. En esta tesis, describimos el uso de DisGeNET para el estudio de las propiedades de los genes asociados a enfermedades en el contexto de redes de interacción entre proteínas. Para ello, evaluamos previamente cómo la utilización de distintos algoritmos de reconocimiento de comunidades en redes afecta a los resultados de los análisis e influencia su interpretación biológica. A continuación, caracterizamos las propiedades de redes de los genes asociados a enfermedades como conjunto y también en sub-grupos, empleando diferentes criterios de clasificaciones de las enfermedades. Posteriormente, evaluamos cómo estas propiedades están relacionadas con la tolerancia a mutaciones posiblemente deletéreas en distintos grupos de genes, mediante el análisis de datos generados por las nuevas tecnologías de secuenciación. Finalmente, desarrollamos una nueva metodología de medicina de sistemas para explorar los mecanismos moleculares de la comorbilidades, y la aplicamos al estudio de las comorbilidades de la enfermedad pulmonar obstructiva crónica

Made available in DSpace on 2014-06-11T19:26:49Z (GMT). No. of bitstreams: 0 Previous issue date: 2011-02-17Bitstream added on 2014-06-13T18:30:37Z : No. of bitstreams: 1 grigoleto_gg_me_sjrp_parcial.pdf: 97076 bytes, checksum: aeb71f1e3dcd408591619b0ee4d085c0 (MD5) Bitstreams deleted on 2015-08-28T16:08:53Z: grigoleto_gg_me_sjrp_parcial.pdf,. Added 1 bitstream(s) on 2015-08-28T16:09:59Z : No. of bitstreams: 1 000641048.pdf: 764505 bytes, checksum: 32067ef8749a30efe912c85f09883a5e (MD5)
O presente trabalho tem por objetivo analisar a tradução de vocábulos considerados recorrentes e preferenciais na obra As mulheres de Tijucopapo, de Marilene Felinto, bem como observar possíveis tendências de normalização por parte da tradutora Irene Matthews na tradução para o inglês: The women of Tijucopapo. Esta pesquisa é parte de um projeto maior, o PETra – Padrões de Estilo de Tradutores, coordenado pela Profa. Dra. Diva Cardoso de Camargo. A metodologia situa-se no campo dos estudos da tradução baseados em corpus (proposta de Baker, 1993, 1995, 1996, 2000, 2004; estudos sobre normalização de Scott, 1998; e pesquisas de Camargo, 2005, 2007), e no da linguística de corpus (estudos de Berber Sardinha, 2003, 2004); também se apoia na fortuna crítica da autora (trabalhos de Araújo, 2006; Bailey, 2010; Fiorucci, 2009; Gonçalves, 2001; Jacomel, 2008; Santos, 2005; Schmidt, 2009; Vieira, 2001; Wanderley, 2009 e Xavier, 2003). A pesquisa foi realizada por meio de uma combinação de análises semimanuais e de análises computadorizadas. Inicialmente, utilizamos o programa WordSmith Tools para obter a frequência das palavras na obra de Felinto. Em seguida, utilizamos a ferramenta KeyWords para identificar quais seriam as palavras-chave presentes no romance. Após esse levantamento, recorremos à fortuna crítica de Felinto e definimos cinco vocábulos considerados preferenciais da autora:homem, mulheres, chuva, amor e égua. Com base em Scott (1998), examinamos a tradução dos cinco vocábulos em relação a nove características de normalização. Os resultados encontrados nesta pesquisa mostram que a tradutora Irene Matthews tende a usar estratégias que podem ser identificadas como características de normalização
This thesis aims at analyzing the translation of words considered recurring and preferred in the novel As mulheres de Tijucopapo, by Marilene Felinto, as well as observing possible tendencies of normalization by the translator Irene Matthews in the translation to English: The women of Tijucopapo. This research is part of a bigger project, the PETra (Padrões de Estilo de Tradutores), coordinated by Doc. Diva Cardoso de Camargo. The methodology employed is that of corpus-based translation studies (proposed by Baker, 1993, 1995, 1996, 2000, 2004; Scott‘s study concerning normalization, 1998; and Camargo‘s research studies, 2005, 2007), and that of corpus linguistics (Berber Sardinha‘s study, 2003, 2004). We also base our study on Felinto‘s critical heritage (studies by Araújo, 2006; Bailey, 2010; Fiorucci, 2009; Gonçalves, 2001; Jacomel, 2008; Santos, 2005; Schmidt, 2009; Vieira, 2001; Wanderley, 2009, and Xavier, 2003). The research was carried out by means of a combination of semi- manual and computerized analyses. Initially, we used the computer software WordSmith Tools to obtain the frequency of the words in Felinto‘s book. Then, we went on to use the tool KeyWords to identify which would be the keywords in the author‘s novel. After collecting these data, we used Felinto‘s critical heritage in order to select five words that could be considered as recurring and preferred in Felinto‘s writing: 'homem‘ ('man‟), 'mulheres‘ ('women)‘, 'chuva‘ ('rain‘), 'amor‘ ('love‘) and 'égua‘ ('mare‘). Based on Scott (1998), we analyzed the translation of the five selected words in relation to nine normalization features. The results obtained in this study show that the translator Irene Matthews tends to use strategies that may be identified as features of normalization

tRNA molecules undergo extensive post-transcriptional modifications to produce several variations of the four canonical nucleotides. Despite the huge number of nucleotide modifications that have been identified in tRNAs, to date their biological roles and regulations continue to be poorly understood. The uridines at the wobble position of the eukaryotic cytoplasmic tRNAs tK^UUU, tQ^UUG and tE^UUC are methoxycarbonylmethylated and thiolated to form mcm⁵s²U₃₄ by the ELP- and URM1-pathways, respectively. Several in vitro experiments have implicated these modifications in modulating wobbling capacity. Moreover, mutations in the ELP- and URM1-pathway genes have been associated with physiological defects in several organisms, including complex neurological disorders in humans. In this thesis we used a systems biology approach to study the in vivo functional relevance of mcm⁵s²U₃₄. A sensitive, robust and quantitative proteomics workflow was developed and applied to investigate differential proteome composition in budding yeast mutants deficient in U₃₄ modifications. We find that, in vivo and under normal conditions, mcm⁵s²U₃₄ fine tunes proteome composition by ensuring efficient translation of mRNAs biased for AAA, CAA and GAA codons. Importantly, our results connect these tRNA modifications with various cellular stress response pathways. Follow up analyses of yeast cells subjected to environmental stresses were conducted and led to the discovery that the biosynthesis of mcm⁵s²U₃₄ is dynamically regulated in response to growth conditions in a URM1-pathway dependent fashion. We propose that this regulation allows the cells to adjust their translational capacity during unfavourable growth conditions and contributes to the management of the environmental stress response. Overall, this thesis presents the first extensive investigation of the functional relevance of tRNA nucleotide modifications and reports one of the few known cases wherein cells regulate the levels of modified nucleotides to fine tune their metabolism in response to environmental cues. We expect that dynamic modulation of RNA modifications will prove to be a more general regulatory mechanism of cellular processes. The experimental and analytical approaches presented in this dissertation will provide a general framework for future studies in this field.

Orientador: Diva Cardoso de Camargo
Banca: Emiliana Fernandes Bonalumi
Banca: Adriane Orenha Ottaiano
Resumo: O presente trabalho tem por objetivo analisar a tradução de vocábulos considerados recorrentes e preferenciais na obra As mulheres de Tijucopapo, de Marilene Felinto, bem como observar possíveis tendências de normalização por parte da tradutora Irene Matthews na tradução para o inglês: The women of Tijucopapo. Esta pesquisa é parte de um projeto maior, o PETra - Padrões de Estilo de Tradutores, coordenado pela Profa. Dra. Diva Cardoso de Camargo. A metodologia situa-se no campo dos estudos da tradução baseados em corpus (proposta de Baker, 1993, 1995, 1996, 2000, 2004; estudos sobre normalização de Scott, 1998; e pesquisas de Camargo, 2005, 2007), e no da linguística de corpus (estudos de Berber Sardinha, 2003, 2004); também se apoia na fortuna crítica da autora (trabalhos de Araújo, 2006; Bailey, 2010; Fiorucci, 2009; Gonçalves, 2001; Jacomel, 2008; Santos, 2005; Schmidt, 2009; Vieira, 2001; Wanderley, 2009 e Xavier, 2003). A pesquisa foi realizada por meio de uma combinação de análises semimanuais e de análises computadorizadas. Inicialmente, utilizamos o programa WordSmith Tools para obter a frequência das palavras na obra de Felinto. Em seguida, utilizamos a ferramenta KeyWords para identificar quais seriam as palavras-chave presentes no romance. Após esse levantamento, recorremos à fortuna crítica de Felinto e definimos cinco vocábulos considerados preferenciais da autora:"homem", "mulheres", "chuva", "amor" e "égua". Com base em Scott (1998), examinamos a tradução dos cinco vocábulos em relação a nove características de normalização. Os resultados encontrados nesta pesquisa mostram que a tradutora Irene Matthews tende a usar estratégias que podem ser identificadas como características de normalização
Abstract: This thesis aims at analyzing the translation of words considered recurring and preferred in the novel As mulheres de Tijucopapo, by Marilene Felinto, as well as observing possible tendencies of normalization by the translator Irene Matthews in the translation to English: The women of Tijucopapo. This research is part of a bigger project, the PETra (Padrões de Estilo de Tradutores), coordinated by Doc. Diva Cardoso de Camargo. The methodology employed is that of corpus-based translation studies (proposed by Baker, 1993, 1995, 1996, 2000, 2004; Scott's study concerning normalization, 1998; and Camargo's research studies, 2005, 2007), and that of corpus linguistics (Berber Sardinha's study, 2003, 2004). We also base our study on Felinto's critical heritage (studies by Araújo, 2006; Bailey, 2010; Fiorucci, 2009; Gonçalves, 2001; Jacomel, 2008; Santos, 2005; Schmidt, 2009; Vieira, 2001; Wanderley, 2009, and Xavier, 2003). The research was carried out by means of a combination of semi- manual and computerized analyses. Initially, we used the computer software WordSmith Tools to obtain the frequency of the words in Felinto's book. Then, we went on to use the tool KeyWords to identify which would be the keywords in the author's novel. After collecting these data, we used Felinto's critical heritage in order to select five words that could be considered as recurring and preferred in Felinto's writing: 'homem' ('man‟), 'mulheres' ('women)', 'chuva' ('rain'), 'amor' ('love') and 'égua' ('mare'). Based on Scott (1998), we analyzed the translation of the five selected words in relation to nine normalization features. The results obtained in this study show that the translator Irene Matthews tends to use strategies that may be identified as features of normalization
Mestre

Le but de cette these est de definir et developper un systeme de bases lexicales multilingues independant des applications et des theories linguistiques. Elle debute donc par une etude des dictionnaires (imprimes ou electroniques) et des outils de gestion de bases lexicales, avec, notamment, le projet ESPRIT MULTILEX, considere comme l'effort le plus avance, mais qui presente quelques faiblesses (architecture lexicale par transfert, structures linguistiques codees sous forme de structures de traits types. ). La deuxieme partie de cette these est consacree a la definition de SUBLIM, un systeme de gestion de bases lexicales multilingues qui permet de specifier l'architecture lexicale (organisation des dictionnaires) et l'architecture linguistique (organisation des informations linguistiques des unites des dictionnaires), sans imposer de contraintes ni sur les types de dictionnaires choisis, ni sur les structures linguistiques utilisees. La troisieme et derniere partie de cette these presente une specialisation de ce systeme generique en un systeme dedie a la gestion de bases lexicales multilingues fondees sur les acceptions interlingues : NADIA. Cette approche generalise certaines methodes interlingues comme celle du projet de traduction multilingue ULTRA, et permet la definition de bases lexicales multilingues ne se basant pas sur une approche par connaissances

Thesis (Ph.D.)--Indiana University, 2009.
Title from screen (viewed on February 2, 2010). Department of Biochemistry and Molecular Biology, Indiana University-Purdue University Indianapolis (IUPUI). Advisor(s): Mark R. Kelley, Millie M. Georgiadis, John J. Turchi, Martin L. Smith. Includes vitae. Includes bibliographical references (leaves 114-133).

Dans cette thèse, nous abordons la réalisation d’un outil innovant d’aide à la traduction anglais/arabe pour répondre au besoin croissant en termes d’outils en ligne d’aide à la traduction centrés sur la langue arabe. Cet outil combine des dictionnaires adaptés aux spécificités de la langue arabe et un concordancier bilingue issu des corpus parallèles. Compte tenu de sa nature agglutinante et non voyellée, le mot arabe nécessite un traitement spécifique. C’est pourquoi, et pour construire nos ressources lexicales, nous nous sommes basés sur l’analyseur morphologique de Buckwalter qui, d’une part, permet une analyse morphologique en tenant compte de la composition complexe du mot arabe (proclitique, préfixe, radical, suffixe, enclitique), et qui, d’autre part, fournit des ressources traductionnelles permettant une réadaptation au sein d’un système de traduction. Par ailleurs, cet analyseur morphologique est compatible avec l’approche définie autour de la base de données DIINAR (DIctionnaire Informatisé de l’Arabe), qui a été construite, entre autres, par des membres de notre équipe de recherche. Pour répondre à la problématique du contexte dans la traduction, un concordancier bilingue a été développé à partir des corpus parallèles Ces derniers représentent une ressource linguistique très intéressante et ayant des usages multiples, en l’occurrence l’aide à la traduction. Nous avons donc étudié de près ces corpus, leurs méthodes d’alignement, et nous avons proposé une approche mixte qui améliore significativement la qualité d’alignement sous-phrastique des corpus parallèles anglais-arabes. Plusieurs technologies informatiques ont été utilisées pour la mise en œuvre de cet outil d’aide à la traduction qui est disponible en ligne (tarjamaan.com), et qui permet à l’utilisateur de chercher la traduction de millions de mots et d’expressions tout en visualisant leurs contextes originaux. Une évaluation de cet outil a été faite en vue de son optimisation et de son élargissement pour prendre en charge d’autres paires de langues
We create an innovative English/Arabic translation aid tool to meet the growing need for online translation tools centered on the Arabic language. This tool combines dictionaries appropriate to the specificities of the Arabic language and a bilingual concordancer derived from parallel corpora. Given its agglutinative and unvoweled nature, Arabic words require specific treatment. For this reason, and to construct our dictionary resources, we base on Buckwalter's morphological analyzer which, on the one hand, allows a morphological analysis taking into account the complex composition of the Arabic word (proclitic, prefix, stem, suffix, enclitic), and on the other hand, provides translational resources enabling rehabilitation in a translation system. Furthermore, this morphological analyzer is compatible with the approach defined around the DIINAR database (DIctionnaire Informatisé de l’Arabe - Computerized Dictionary for Arabic), which was constructed, among others, by members of our research team. In response to the contextual issue in translation, a bilingual concordancer was developed from parallel corpora. The latter represent a novel linguistic resource with multiple uses, in this case aid for translation. We therefore closely analyse these corpora, their alignment methods, and we proposed a mixed approach that significantly improves the quality of sub-sentential alignment of English-Arabic corpora. Several technologies have been used for the implementation of this translation aid tool which have been made available online (tarjamaan.com) and which allow the user to search the translation of millions of words and expressions while visualizing their original contexts. An evaluation of this tool has been made with a view to its optimization and its enlargement to support other language pairs

Afin de limiter les dégradations liées au codage bas-débit des signaux audionumériques, la stratégie adoptée par la plupart des systèmes de compression de parole et de musique consiste à ne pas transmettre le contenu hautes-fréquences. C'est ainsi qu'aux environs des 20kbit/s, les codeurs de musique actuels ne restituent pas les sons avec leur qualité naturelle (leur bande passante étant limitée aux environs des 6kHz). Les sons ainsi codés/décodés deviennent ternes et perdent de leur qualité. On se propose d'étudier dans cette thèse de nouvelles techniques susceptibles de palier à cette perte des aigus. Les systèmes d'enrichissement de spectre permettent, avec très peu de données additionnelles, de rehausser la bande passante, et donc la qualité de ces signaux à bande-limitée. Le principe de ces techniques consiste à exploiter les informations comprises dans le spectre basse-fréquence afin de synthétiser le signal pleine-bande de qualité proche de celle de l'original. Dans le cadre d'un contrat financé par France Télécom R&D , l'objectif de cette thèse est la réalisation d'un système d'enrichissement de spectre des signaux audionumériques (parole et musique). La technique PAT (Perceptual Audio Transposition) implémentée a fait l'objet de deux propositions de normalisation dans les instances DRM (Digital Radio Mondiale) et MPEG-4 (Moving Picture Experts Group). Le document est structuré en 4 parties. La première partie s'attache à introduire les principes de l'extension de bande en se fondant sur les propriétés psychoacoustiques et les caractéristiques des signaux audio mis en jeux. Grâce à cette étude préalable, l'enrichissement des signaux sonores est réalisé en deux étapes: une étape d'extension de la structure fine du spectre et une étape d'ajustement de l'enveloppe, qui font l'objet des deux chapitres suivants. Ainsi, la seconde partie est consacrée aux techniques d'estimation, de transmission et d'ajustement d'enveloppe spectrale. Deux techniques particulières sont développées: L'une basée sur la prédiction linéaire et l'autre sur la modélisation d'enveloppe par facteurs d'échelle dans le domaine fréquentiel. Dans la troisième partie, les différentes solutions permettant d'étendre la structure fine spectrale sont abordées. L'étude s'est portée notamment sur les translations de spectre dans le domaine fréquentiel et sur les distorsions non-linéaires. Enfin, en quatrième partie, on présente un schéma complet d'enrichissement de spectre avant d'en évaluer ses performances dans le cadre de la normalisation MPEG-4. Une toute nouvelle technique de compression des signaux audionumériques est ainsi introduite dans cette thèse. Celle-ci a montré un réel intérêt dans le domaine de la compression du son. Pour une qualité équivalente, la réduction de débit obtenue est de l'ordre de 25%.

Die vorliegende Dissertationsschrift untersucht die deutschen verbonominalen Konstruktionen vom Typ Verb + (Präposition) + Substantiv und ihre Entsprechungen im Litauischen. Da diese Untersuchung eine kontrastive Studie ist, ist das Ziel der Arbeit die kontrastive Untersuchung und die Beschreibung der möglichen Übersetzungsprobleme sowie die Suche nach den passenden Lösungen. Bei dieser Studie soll kontrastiv vorgegangen werden, deswegen werden die beschreibende und die kontrastive Untersuchungsmethode herangezogen. Es wird aber auch der korpuslinguistischen Methode bedient, weil die meisten jüngsten Kollokationsuntersuchungen darauf basieren. In dieser Arbeit werden auch verschiedene qualitative und quantitative Parameter verglichen. Die deutschen verbonominalen Konstruktionen und ihre möglichen litauischen Entsprechungen wurden in der vorliegenden Arbeit auf zwei Ebenen untersucht: auf der formalen oder strukturellen und auf der Bedeutungs- oder semantischen Ebene. Für die durchgeführte Analyse wurden 1230 Beispiele mit den deutschen verbonominalen Konstruktionen und ihre Übersetzungen ins Litauische herangezogen. Im praktischen Teil der vorliegenden Dissertationsschrift werden die konkreten Übersetzungsbeispiele analysiert und die möglichen Problemlösungen vorgeschlagen. Da die Übersetzer, die mit dem Sprachpaar Deutsch-Litauisch arbeiten, nur das einzige zweisprachige Wörterbuch konsultieren können, wird es näher hinsichtlich der Informationen zu den stabilen... [der volle Text, siehe weiter]
Šis tyrimas skirtas kolokacijų vertimo problemoms. Darbo objektas – vokiečių kalbos verbonominalinės konstrukcijos, t. y. veiksmažodis + (prielinksnis) + daiktavardis junginiai. Jo tikslas – gretinant dvi kalbas, aptarti ir susisteminti kylančias vertimo problemas bei ieškoti būdų joms spręsti. Darbe taikomi aprašomasis, gretinamasis ir tekstynų lingvistikos tyrimo metodai bei lyginami įvairūs kiekybiniai ir kokybiniai duomenys. Disertacijos tiriamąją medžiagą sudaro apie 4 tūkst. konstrukcijų. Tyrimui sukauptas ir tirtas apie 1 mln. žodžių lygiagretusis tekstynas, sudarytas iš grožinės, populiariosios mokslinės literatūros, kanceliarinių bei administracinių tekstų. Tekstyne rasta 651 skirtinga konstrukcija, o bendras verbonominalinių konstrukcijų vartojimo atvejų skaičius – 1230. Disertacijoje aptariami konkretūs tiriamų konstrukcijų vertimo atvejai, nagrinėjamos formos ir reikšmės lygmenyse vertėjams kylančios problemos ir siūlomi galimi jų sprendimo būdai. Kadangi iš/į vokiečių kalbą verčiantys vertėjai priversti naudotis vieninteliu dvikalbiu vokiečių–lietuvių kalbų žodynu, darbe analizuojama, kaip jame pateikiami stabilieji junginiai, bei konstatuojama, kad informacija pateikiama nesistemingai, todėl pristatomas dvikalbės internetinės verbonominalinių konstrukcijų duomenų bazės modelis.
This study deals with problems of translation of collocations between German and Lithuanian. The syntactic pattern under analysis is Verb + (Preposition) + Noun, i.e. so-called verbal-nominal constructions in German. The aim of the study is to contrast the two languages and identify translation problems as well as propose appropriate solutions. Methodology used for the analysis involves descriptive, contrastive and corpus linguistics approaches, including processing of various quantitative and qualitative data. The material analysed in the dissertation consists of ca. 4,000 constructions. A parallel corpus of 1 million words was used. It was compiled from texts of fiction, popular science writing and administrative office communication. 651 construction types of the chosen pattern were identified in the corpus with the total frequency of tokens of 1230 items. The analysis was focused on the translation of individual types of constructions which often pose a number of difficulties at the level of form and meaning. Owing to the fact that translators between German and Lithuanian have only one German-Lithuanian dictionary at their disposal, the study also involves a critical revision of this lexicographic tool which, as it was found, often contains inadequate information on collocations. The author proposes a model of a bilingual electronic data base as a more advanced tool for the translation of verbal-nominal constructions.

A partir des années 1960, la fiction hispano-américaine connaît un essor sans précédent, auquel la critique littéraire internationale a donné le nom très contesté de boom. Le principal problème de ce terme est qu’il renvoie à un phénomène touchant à la fois la production et la diffusion de la littérature hispano-américaine.Notre travail prose une relecture du phénomène à travers l’étude de la traduction et de la réception de Julio Cortázar, Carlos Fuentes, Gabriel García Márquez et Mario Vargas Llosa en France et aux Pays-Bas entre 1960 et 1973.En retraçant l’historique des premières traductions et en analysant le contexte éditorial dans lequel elles sont arrivées, nous montrons l’existence d’écarts profonds séparant les deux pays dans la traduction des quatre auteurs, n’illustrant pas toujours leur position théorique au sein de l’espace mondial de la traduction, et le rôle importants joués par certains intermédiaires culturels. Par le biais d’une étude de traduction et de réception, nous mettons aussi en évidence le rôle de médiation joué par les traducteurs et la critique littéraire entre les œuvres originales et le lecteur final dans les deux pays de référence
Spanish-American fiction has been developing unprecedented from the 1960’s, what has been given the highly contested name of boom by international literary critics. The main problem of that term is that it refers to a phenomenon affecting both the production and distribution of Spanish-American literature.Our work is a reading of the phenomenon through the study of translation and reception of Julio Cortázar, Carlos Fuentes, Gabriel García Márquez and Mario Vargas Llosa in France and in the Netherlands between 1960 and 1973.In tracing the history of the early translations and analysis of the editorial context in which they arrived, we show the existence of profound differences between the two countries in the translation of the four authors, which do not illustrate their theoretical position in the world space of translation, and the important role played by certain cultural intermediaries. Through a study of translation and reception, we also highlight the mediating role played by translators and literary criticism between the original texts and the final reader in the two reference countries

Šis tyrimas skirtas kolokacijų vertimo problemoms. Darbo objektas – vokiečių kalbos verbonominalinės konstrukcijos, t. y. veiksmažodis + (prielinksnis) + daiktavardis junginiai. Jo tikslas – gretinant dvi kalbas, aptarti ir susisteminti kylančias vertimo problemas bei ieškoti būdų joms spręsti. Darbe taikomi aprašomasis, gretinamasis ir tekstynų lingvistikos tyrimo metodai bei lyginami įvairūs kiekybiniai ir kokybiniai duomenys. Disertacijos tiriamąją medžiagą sudaro apie 4 tūkst. konstrukcijų. Tyrimui sukauptas ir tirtas apie 1 mln. žodžių lygiagretusis tekstynas, sudarytas iš grožinės, populiariosios mokslinės literatūros, kanceliarinių bei administracinių tekstų. Tekstyne rasta 651 skirtinga konstrukcija, o bendras verbonominalinių konstrukcijų vartojimo atvejų skaičius – 1230. Disertacijoje aptariami konkretūs tiriamų konstrukcijų vertimo atvejai, nagrinėjamos formos ir reikšmės lygmenyse vertėjams kylančios problemos ir siūlomi galimi jų sprendimo būdai. Kadangi iš/į vokiečių kalbą verčiantys vertėjai priversti naudotis vieninteliu dvikalbiu vokiečių–lietuvių kalbų žodynu, darbe analizuojama, kaip jame pateikiami stabilieji junginiai, bei konstatuojama, kad informacija pateikiama nesistemingai, todėl pristatomas dvikalbės internetinės verbonominalinių konstrukcijų duomenų bazės modelis.
This study deals with problems of translation of collocations between German and Lithuanian. The syntactic pattern under analysis is Verb + (Preposition) + Noun, i.e. so-called verbal-nominal constructions in German. The aim of the study is to contrast the two languages and identify translation problems as well as propose appropriate solutions. Methodology used for the analysis involves descriptive, contrastive and corpus linguistics approaches, including processing of various quantitative and qualitative data. The material analysed in the dissertation consists of ca. 4,000 constructions. A parallel corpus of 1 million words was used. It was compiled from texts of fiction, popular science writing and administrative office communication. 651 construction types of the chosen pattern were identified in the corpus with the total frequency of tokens of 1230 items. The analysis was focused on the translation of individual types of constructions which often pose a number of difficulties at the level of form and meaning. Owing to the fact that translators between German and Lithuanian have only one German-Lithuanian dictionary at their disposal, the study also involves a critical revision of this lexicographic tool which, as it was found, often contains inadequate information on collocations. The author proposes a model of a bilingual electronic data base as a more advanced tool for the translation of verbal-nominal constructions.
Die vorliegende Dissertationsschrift untersucht die deutschen verbonominalen Konstruktionen vom Typ Verb + (Präposition) + Substantiv und ihre Entsprechungen im Litauischen. Da diese Untersuchung eine kontrastive Studie ist, ist das Ziel der Arbeit die kontrastive Untersuchung und die Beschreibung der möglichen Übersetzungsprobleme sowie die Suche nach den passenden Lösungen. Die deutschen verbonominalen Konstruktionen und ihre möglichen litauischen Entsprechungen wurden in der vorliegenden Arbeit auf zwei Ebenen untersucht: auf der formalen oder strukturellen und auf der Bedeutungs- oder semantischen Ebene. Für die durchgeführte Analyse wurden 1230 Beispiele mit den deutschen verbonominalen Konstruktionen und ihre Übersetzungen ins Litauische herangezogen. Da die Übersetzer, die mit dem Sprachpaar Deutsch-Litauisch arbeiten, nur das einzige zweisprachige Wörterbuch konsultieren können, wird es näher hinsichtlich der Informationen zu den stabilen Wortverbindungen untersucht und festgestellt, dass eine zweisprachige Online-Datenbank beim Übersetzen der verbonominalen Konstruktionen hilfreicher wäre.

Le travail présenté dans ce mémoire concerne la modélisation système des circuits conformant têtes d'émission et réception de télécommunications. Le premier chapitre sert à mettre en évidence le besoin des modèles réduits, capables de caractériser de forme précise le comportement des circuits et systèmes, ainsi qu'à réviser l'état de l'art de ce domaine. Le deuxième chapitre est consacré aux méthodes de réduction de modèles pour des circuits pompés et non pompés, qui présentent un comportement entrée-sortie linéaire avec mémoire. Ensuite, une technique de réduction pour des circuits pompés, faiblement non linéaires avec mémoire est proposée dans le troisième chapitre. Dans le quatrième chapitre, les méthodes développées sont appliquées à la modélisation d'une chaîne d'émission et de réception. Finalement, les conclusions et perspectives de ce travail sont présentées au dernier chapitre
The work reported deals with system-level modelling techniques of microwave circuits commonly employed in emission and reception telecommunications channels. The need of reduced models for simulation of communication systems is discussed, and the state of the art of this activity is summarized in the first chapter. The second chapter focus on model reduction methods for pumped and non-pumped circuits exhibiting a linear input-output behaviour. Then, in the third chapter, a model reduction technique for pumped weakly non linear circuits with memory is presented. In the fourth chapter, the proposed methods are applied to the modelling and simulation of a transmitter and receiver. Finally, the conclusions and perspectives of this work are summarised

Presentation du langage pascal/relationnel construit a partir de pascal par adjonction de constructions venant du domaine des bases de donnees ainsi que la conception et l'implantation du traducteur du langage. Les instructions relationnelles sont inspirees du calcul des predicats et permettent les operations classiques offertes par les systemes de gestion de bases de donnees relationnelles

Les services de stérilisation sont des lieux de production de soins caractérisés par une multitude d’activités et situations auxquelles ils sont confrontés. En outre, les services de stérilisation doivent assurer leurs missions dans un environnement caractérisé par la présence d’une variété de risques. Les risques présents dans ces milieux peuvent aller des pannes des équipements jusqu’aux contaminations et transmission des maladies nosocomiales. Ces services sont aussi tenus de garder un niveau de performances satisfaisant pour assurer la continuité des soins dans les blocs opératoires.Pour aider ces services dans leur quête d'un système performant, capable d’évoluer dans un environnement à haut niveau de risques, nous nous intéressons dans ce travail de recherche au développement d’une approche intégrée pour l’analyse de risques et l’évaluation des performances. Ce travail s’intègre dans un cadre collaboratif entre le laboratoire G-SCOP et le service de stérilisation du CHU de Grenoble, terrain d’étude choisi pour mettre en œuvre l’approche proposée.L'approche que nous proposons se déroule en plusieurs étapes: tout d’abord, suite à une comparaison entre les méthodes de gestion des risques, nous nous sommes orientés vers l’approche pilotée par modèle, dénommée FIS (Fonction Interaction Structure). En nous basant sur FIS, nous avons développé un modèle de risque dans ce service de stérilisation, décrivant à la fois les fonctions, les ressources permettant la réalisation de ces fonctions ainsi que les différents risques qui peuvent être rencontrés. Dans un deuxième temps, nous avons représenté le comportement dynamique du modèle de risques obtenu. Ce modèle dynamique permet de simuler le comportement du service de stérilisation et le voir évoluer dans les situations normales de fonctionnement et les situations de risques. Pour ce faire, nous avons introduit une nouvelle classe de réseau de Petri appelée réseau de Petri PTPS (Predicate-Transition, Prioritized, Synchronous) permettant de représenter et simuler le comportement dynamique du modèle FIS. Par la suite, nous avons automatisé le passage entre le modèle de risque et le modèle dynamique. Cette automatisation est effectuée par un ensemble d’algorithmes de traduction, capables de convertir automatiquement le modèle FIS et le modèle de simulation en réseau de Petri PTPS.Cette approche a donné lieu à un outil de modélisation et de simulation en mode dégradé, appelé SIM-RISK. Nous avons également montré l’utilité de cet outil sur des exemples inspirés des différents risques rencontrés dans le service de stérilisation
Sterilization services are vulnerable to risks, due to the contagious nature of their environment and to the degradation that risks can cause to their performances and to the safety of patients and staff. The risks in these facilities can range from equipment failure to the transmission of nosocomial infections and diseases. In this kind of high risk environment, these services are also required to maintain an adequate level of performance to ensure continuity of care in operating theaters.We focus in this research on the development of an integrated approach for risk analysis and performance assessment. This work is part of a collaborative work between the G-SCOP laboratory and the sterilization service of the University Hospital of Grenoble, which was the case study chosen to implement the proposed approach.The approach we propose is conducted in several steps: first, following a comparison of the risk analysis methods, we have chosen a model driven approach called FIS (Function Interaction Structure). Based on FIS, we have developed a risk model of Grenoble University Hospital sterilization service. This model describes the functions, the resources to achieve these functions as well as the various risks that may be encountered. Secondly, we introduced a new view to the FIS model dedicated to describe the dynamic behaviour of the resulting risk model.This dynamic model can simulate the behaviour of the sterilization service in normal situations of operations and risk situations.To do this, we have introduced a new Petri Net class called PTPS (Predicate-Transition, Prioritized, Synchronous) Petri Net to represent and simulate the dynamic behaviour of the FIS model. Subsequently, we automated the switching between the risk model and the dynamic model. This automation is performed by a set of translation algorithms capable of automatically converting the FIS model to a PTPS Petri Net simulation model .This approach resulted in a modelling and simulation tool in degraded mode called SIM-RISK. We also showed the usefulness of this tool by some examples based on different risks encountered in the sterilization service

Un premier objectif de cette thèse est la constitution de ressources linguistiques pour un système de traduction automatique statistique factorisée français - roumain. Un deuxième objectif est l’étude de l’impact des informations linguistiques exploitées dans le processus d’alignement lexical et de traduction. Cette étude est motivée, d’une part, par le manque de systèmes de traduction automatique pour la paire de langues étudiées et, d’autre part, par le nombre important d’erreurs générées par les systèmes de traduction automatique actuels. Les ressources linguistiques requises par ce système sont des corpus parallèles alignés au niveau propositionnel et lexical. Ces corpus sont également segmentés lexicalement, lemmatisés et étiquetés au niveau morphosyntaxique
Our first aim is to build linguistic resources for a French - Romanian factored phrase - based statistical machine translation system. Our second aim is to study the impact of exploited linguistic information in the lexical alignment and translation process. On the one hand, this study is motivated by the lack of such systems for the studied languages. On the other hand, it is motivated by the high number of errors provided by the current machine translation systems. The linguistic resources required by the system are tokenized, lemmatized, tagged, word, and sentence - aligned parallel corpora

Made available in DSpace on 2016-06-02T19:05:53Z (GMT). No. of bitstreams: 1 3952.pdf: 2039842 bytes, checksum: 13562902fbbd07a996facebf66d35dd7 (MD5) Previous issue date: 2011-10-21
Financiadora de Estudos e Projetos
In this research, we describe how common sense knowledge with machine translation can help the communication among people with different cultural background. In order to evaluate this possibility, we developed a bilingual chat called Culture-to-Chat or C2C that provides a communication channel and has resources that help its user to create messages in a second language. In the computer-mediated communication field, it is possible to notice that people are crossing geographic borders and having opportunities to share experiences among different cultures. Sharing information in a non native language can be difficult to some users. Some computational tools that support communication uses machine translation to aid users that need to work with different language. C2C also uses this approach and adopts a semantic network of cultural knowledge, collaboratively built on the Web through the Open Mind Common Sense in Brazil project (OMCS-Br) to work with cultural expression, in other words, terms whose meaning depends on user‟s culture. Following a user-centered design approach that focuses on prototyping, we present the development of C2C passing by low, middle and high fidelity prototypes. In order to observe how this computational tool is used and collect the opinion of target users, we perform a pilot study involving Brazilian and Canadian users. This study showed some enhancements for the tool and pointed evidences that this chat contributes to the communication between people with different cultural backgrounds.
Neste trabalho é descrito como o conhecimento de senso comum em conjunto com a tradução automática pode apoiar a comunicação entre pessoas de diferentes culturas. Para verificar a viabilidade do senso comum foi desenvolvido um chat bilíngue chamado Culture-to-Chat ou C2C, que, além de prover um canal de comunicação, possui mecanismos que auxiliam o usuário na criação de mensagens em língua estrangeira. No campo da comunicação mediada por computador, percebe-se que as pessoas estão cruzando as fronteiras geográficas e tendo oportunidades de troca de experiências entre diferentes culturas. Para alguns usuários, essa troca de informações é feita em uma língua não nativa, o que pode ser uma tarefa difícil para eles. Algumas ferramentas computacionais que apoiam a comunicação propõem o uso de tradução automática para trabalhar com usuários falantes de diferentes línguas. O C2C, além de adotar tal estratégia, adota a rede semântica de conhecimento cultural construida colaborativamente via Web do projeto Open Mind Common Sense no Brasil (OMCS-Br) para trabalhar com expressões culturais, ou seja, termos cujo significado depende da cultura do usuário. Utilizando uma abordagem centrada no usuário focando-se em prototipação, o desenvolvimento do C2C e suas funcionalidades são apresentados por meio de protótipos de diferentes níveis de fidelidade (baixa, média e alta). Com o intuito de observar o uso dessa ferramenta computacional e coletar opiniões de usuários, foi realizado um estudo piloto envolvendo usuários brasileiros e canadenses. Tal estudo mostrou possíveis melhorias pra a ferramenta e apontou indícios de que este chat contribui na comunicação entre pessoas de diferentes culturas.

Bien qu’ils soient largement utilisés dans le domaine, les mesures vibratoires classiques présentent plusieurs limites. A la base, l’analyse vibratoire ne peut identifier qu’environ 60% des défauts qui peuvent survenir dans les machines. Cependant, les principaux inconvénients des mesures de la vibration sont l’accès difficile au système de transmission afin d’y placer le capteur ainsi que le coût conséquent de la mise en œuvre. Ceci résulte en des problèmes de sensibilité relatifs à la position de l’installation et ceux de difficulté pour distinguer la source de vibration à cause de la diversité des excitations mécaniques qui existent dans l’environnement industriel.Par conséquent, l’analyse des signatures du courant électrique des moteurs s’impose comme une alternative prometteuse à l’analyse vibratoire et a donc fait l’objet d’une attention grandissante au cours des dernières années. En effet, l’analyse des signatures électriques a l’avantage d’être une méthode techniquement accessible, non-intrusive au système et peu coûteuse. Les techniques basées sur le courant et la tension ne requièrent que les mesures électriques du moteur qui sont souvent déjà surveillées pour le contrôle et la protection des machines électriques. Ce processus a été principalement utilisé pour la détection des défauts de moteur tels que la rupture de barres du rotor et les défauts d’excentricité ainsi que les défauts de roulements. En revanche, très peu de recherches concernent la détection des défauts en utilisant l’analyse du courant. En outre, les signaux électriques triphasés sont caractérisés par des représentations géométriques particulières liées à leur forme d’onde qui peuvent servir en tant qu’indicateurs différents offrant des informations supplémentaires. Parmi ces indicateurs géométriques, les transformées de Park et de Concordia modélisent les composantes électriques dans un repère bidimensionnel et toute déviation par rapport à la représentation d’origine indique l’apparition d’un dysfonctionnement. Aussi, les équations différentielles de Frenet-Serret représentent la trajectoire du signal dans un espace euclidien tridimensionnel et indiquent ainsi tout changement dans l’état du système. Bien qu’ils aient été utilisés pour les défauts de roulements, ces indicateurs n’ont pas été appliqués dans la détection des défauts d’engrenages en utilisant l’analyse des signatures des courants électriques. D’où l’idée novatrice de combiner ces indicateurs avec des techniques de traitement de signal, ainsi que des techniques de classification pour le diagnostic des engrenages en utilisant l’analyse des signatures de courant et de tension du moteur électrique.Ainsi, dans ce travail, on propose une nouvelle approche pour le diagnostic des défauts d’engrenages en utilisant l’analyse des courants et des tensions électriques du stator de la machine et ceci en se basant sur un ensemble d’indicateurs géométriques (Transformées de Park et de Concordia ainsi que les propriétés du repère Frenet-Serret). Ces indicateurs font partie d’une bibliothèque de signatures de défauts qui a été construite et qui comprend également les indicateurs classiques utilisés pour un large éventail de défauts. Ainsi, un algorithme combine les acquisitions expérimentales des signaux électriques à des méthodes de traitement de signal avancées (décomposition modale empirique,…). Ensuite, celui-ci sélectionne les indicateurs les plus pertinents au sein de la bibliothèque en se basant sur les algorithmes de sélection de paramètres (sélection séquentielle rétrograde et analyse des composantes principales). Enfin, cette sélection est utilisée pour la classification non-supervisée (K-moyennes) pour la distinction entre l’état sain et l’état défaillant
Although they are widely used, classical vibration measurements have several limitations. Vibration analysis can only identify about 60% of the defects that may occur in mechanical systems. However, the main drawbacks of vibration measurements are the difficult access to the transmission system in order to place the sensor as well as the consequent cost of implementation. This results in sensitivity problems relative to the position of the installation and the difficulty to distinguish the source of vibration because of the diversity of mechanical excitations that exist in the industrial environment.Hence, the Motor Current Signatures Analysis (M.C.S.A.) represents a promising alternative to the vibration analysis and has therefore been the subject of increasing attention in recent years. Indeed, the analysis of electrical signatures has the advantage of being a technically accessible method as well as inexpensive and non-intrusive to the system. Techniques based on currents and voltages only require the motor’s electrical measurements which are often already supervised for the purposes of the control and the protection of the electrical machines. This process was mainly used for the detection of motors faults such as rotor bars breakage and eccentricity faults as well as bearings defects. On the other hand, very little research has been focused on gear faults detection using the current analysis. In addition, three-phase electrical signals are characterized by specific geometric representations related to their waveforms and they can serve as different indicators providing additional information. Among these geometric indicators, the Park and Concordia transforms model the electrical components in a two-dimensional coordinate system and any deviation from the original representation indicates the apparition of a malfunction. Moreover, the differential equations of Frenet-Serret represent the trajectory of the signal in a three-dimensional euclidean space and thus indicate any changes in the state of the system. Although they have been previously used for bearing defects, these indicators have not been applied in the detection of gear defects using the analysis of electrical current signatures. Hence, the innovative idea of combining these indicators with signal processing techniques, as well as classification techniques for gears diagnosis using the three-phase motor’s electrical current signatures analysis is established.Hence, in this work, a new approach is proposed for gear faults diagnosis using the motor currents analysis, based on a set of geometric indicators (Park and Concordia transforms as well as the properties of the Frenet-Serret frame). These indicators are part of a specifically built fault signatures library and which also includes the classical indicators used for a wide range of faults. Thus, a proposed estimation algorithm combines experimental measurements of electrical signals with advanced signal processing methods (Empirical Mode Decomposition, ...). Next, it selects the most relevant indicators within the library based on feature selection algorithms (Sequential Backward Selection and Principal Component Analysis). Finally, this selection is combined with non-supervised classification (K-means) for the distinction between the healthy state and faulty states. It was finally validated with a an additional experimental configuration in different cases with gear faults, bearing faults and combined faults with various load levels

En 1204, le roi d’Aragon devint seigneur de Montpellier. À partir de 1309, la curie apostolique s’installa à Avignon. Sous ces deux patronages, l’Université de médecine connut, de la seconde moitié du XIIIe siècle à la fin du XIVe, sa période la plus florissante. Ce fut la période de plus grande activité littéraire de ses gradués. Après avoir assimilé les œuvres arabes traduites au Mont-Cassin et à Tolède, les maîtres montpelliérains étaient à la recherche de nouveaux textes, tant pour leur enseignement que pour leur exercice, notamment au sein des deux cours. Or la récente intégration à la Couronne d’Aragon des territoires du Levante donnait accès à de nouveaux textes issus d’al-Andalus. Sous l’impulsion des maîtres montpelliérains, prospéra ainsi un second foyer de traduction « espagnol ». L’étude des citations faites par les auteurs médicaux montpelliérains montre que sept œuvres andalouses, traduites en latin entre le XIIIe et le XIVe siècles, sont citées dans une vingtaine d’œuvres montpelliéraines. Certaines d’entre elles, qui datent des XVe et XVIe siècles, prouvent que les textes traduits sont restés pendant des décennies à la disposition des étudiants montpelliérains. Ces traductions ne fournirent qu’une faible proportion des citations d’auteurs andalous trouvées dans les textes montpelliérains. Mais elles associent à l’activité du studium montpelliérain de nouveaux auteurs andalous. Le recours à ceux-ci témoigne du dynamisme de l’Université de médecine au Moyen Âge, qui conduisit ses maîtres vers la terra incognita de nouvelles œuvres andalouses
In 1204, the King of Aragon became Lord of Montpellier. In 1309, the Curia apostolic was installed in Avignon. Under these two patronages, the University of medicine experienced, from the second half of the thirteen century to the end of the fourteenth century, its most flourishing period. It was the period of the greatest literary activity of its graduates. After they had digested the Arabic works translated at Montecassino and Toledo, the Montpellier masters were looking for new texts in order to teach as well as to practise, especially within the two courts. Now the recent integration of the territories of Levante into the Crown of Aragon gave access to new texts from al-Andalus. So, on the impulse of the Montpellier masters, translations flourished in a second « Spanish » place. The study of the quotations made by the medical authors of Montpellier show that seven Andalusian works, translated into latin between the thirteen century and the fourteenth century, are quoted in about twenty works from Montpellier. Some among them, dating from the fifteenth and sixteenth centuries, prove that the translated texts remained at the disposal of the Montpellier students for decades. These translations make up only a small proportion of the quotations of Andalusian authors found in the Montpellier texts. But they associate some new Andalusian authors with the activity of the Montpellier studium. The resort to these authors attest the dynamism of the medical university in the Middle Ages which led its masters to the terra incognita of new Andalusian works
En 1204, el Rey de Aragón se volvió Señor de Montpellier. A partir de 1309, la curia apostólica se instaló en Aviñón. Bajo aquellos dos patrocinios, la Universidad de medicina conoció, desde la segunda mitad del siglo XIII hasta el final del siglo XIV, su periódo más próspero. Fue el periódo de la más gran actividad literaria de sus graduados.Despues de haber asimilado las obras árabes traducidas en el Montecassino y en Toledo, los maestros de Montpellier estaban buscando nuevos textos, tanto para su enseñanza como para su desempeño, particularmente al seno de las dos cortes. Pues la reciente integración de los territorios del Levante a la Corona de Aragón daba el acceso a nuevos textos que provenían de al-Andalus. Bajo el impulso de los maestros de Montpellier, prosperó así un segundo foco « español » de traducción. El estudio de las citas hechas por los autores médicos de Montpellier demuestra que siete obras andalusíes, traducidas al latín entre el siglo XIII y el siglo XIV, están citadas en unas veinte obras de Montpellier. Algunas de aquellas, que datan de los siglos XV y XVI, prueban que los textos traducidos quedaron durante decenios a la disposición de los estudiantes de Montpellier. Aquellas traducciones no produjeron más que una escasa proporción de las citas de autores andalusíes encontradas en los textos de Montpellier. Pero asocian nuevos autores andalusíes a la actividad del studium de Montpellier. El recurso a aquellos da un testimonio del dinamismo de la Universidad de medicina en la Edad Media, que llevó sus maestros hacia la terra incognita de nuevas obras andalucíes

This essay deals with translation changes and problems that occur when translating collocations, i.e. sequences of words that are usually combined such as work hard, from Swedish to German. Being an essential part of our communication, collocations are common both in everyday-life and in technical terminology. This qualitative research deals with collocations in scientific texts about children’s literature and aims to increase attention on collocations in the translation process. The term collocation as used here refers to the concept of hierarchical relations between the lexemes of a collocation by the German linguist F. J. Hausmann (1984), distinguishing between bases and collocators. It proved itself useful in the translation process and showed that the translator must pay special attention to the collocator in order to find the equivalent expression in the target language. In doing so he/she is not at all obliged to translate a collocation with the respective target collocation. In fact due to stylistic reasons collocations were replaced by individual lexemes or paraphrases. Referring to Hausmann’s terminology and the studies of Baker (1992) this essay highlights the importance of considering collocations as translation units, i.e. not translating the lexemes of a collocation separately. Furthermore the assumption was disproved that translation errors, e.g. interferences and semantic deficits, are mainly caused when translating collocations to a foreign language. Even translators translating into their mother tongue have to examine collocations carefully to avoid errors that can result in doubts about the quality of the whole translation.

One of the key goals of Natural Language Processing is to make computers understand natural language. Semantic Parsing has been one of the driving tasks for Natural Language Understanding. It is formally defined as the task of generating meaning representation from natural language input. In this work, we focus on using the Bash command as the meaning representation. Bash is a Unix command language used for interacting with the Operating System. Recent works on natural language to Bash command translation have made significant advances on this problem. The best performing solutions employ a neural network architecture called the Transformer. In this work, we explore the aspects of explainability and efficiency for this task and use the Transformer as one of the baselines for comparing the proposed approaches. In the first part, we utilize documentation data from Linux manual pages and the Abstract Syntax Tree for Bash to generate explanations for the translated Bash command. We propose a novel architecture that incorporates tree structure information in the Transformer and provides explanations for its predictions via alignment matrices between user invocation and manual page text. We find that the proposed method performs on par with the Transformer performance. Our method performs better than fine-tuned T5, a Transformer-based neural model pre-trained on a large amount of text data in a self-supervised manner. In the second part, we use the problems inherent synchronous structure and propose the Segmented Invocation Transformer (SIT) that utilizes the information from the constituency parse tree of the natural language invocation. Our method is motivated by the alignment between segments in the natural language text and Bash command components. By utilizing this structure, the proposed method outperforms the state-of-the-art approach while achieving a 1.8x improvement in the inference time (as measured on a CPU) and a 5x reduction in model parameters. We also conduct an attribution analysis using Integrated Gradients to empirically confirm the identified structure and the ability of SIT to capture it.

碩士
國立成功大學
資訊工程學系碩博士班
100
An automatic image analysis system for the diagnosis of knee cruciate ligament is presented. This system segments objects which are used to define some feature points and parameters for cruciate ligament evaluation. Other than some conventional parameters given in literature, we also design new parameters which can be used for cruciate ligament diagnosis. The current cruciate ligament diagnosis is commonly done by doctor’s manual measurement based on the Magnetic Resonance images. Therefore, the diagnostic accuracy varies from person to person. Because of the poor contrast and the complex structure of the knee, we apply a model-based segmentation algorithm to segment out objects of interest from the lateral view of knee radiograph. The appearance model and the shape model of objects of interest in the image are established and used for segmentation. By using the proposed model-based algorithm, the contour which best matches with the model can be successfully segmented. For the complicated femoral condyles, we use a dual condyles model which simultaneously segments two condyles using the relationship between the condyles. In the experiments, the diagnostic accuracy of each measurement is promising. Furthermore, several factors that affect the reliability of measurements are also analyzed.

Relatório de estágio de mestrado em Tradução e Comunicação Multilingue
Um estágio curricular é uma oportunidade para colocar em prática os conhecimentos adquiridos durante o decorrer do mestrado e serve como uma primeira experiência profissional no mundo dos serviços linguísticos. Ao fim de três meses, aprendi o que implica ser um tradutor dentro de uma empresa, os diferentes passos do processo de tradução, como utilizar as ferramentas informáticas de apoio à tradução utilizadas pelos profissionais e técnicas de pesquisa. A importância do uso de bases terminológicas na tradução técnica tornou-se evidente nos 45 projetos realizados no decorrer do estágio, cada qual com as suas particularidades linguísticas e terminologia específica da respetiva área de trabalho. Todas as ferramentas de apoio à tradução utilizadas durante o estágio trabalhavam com bases terminológicas e memórias de tradução e foi possível, através das mesmas, melhorar a eficácia e precisão das competências de tradução do estagiário. O uso de CAT-Tools é, hoje em dia, absolutamente indispensável para qualquer tradutor que se dedique à tradução técnica.
A curricular internship is an opportunity to put the knowledge acquired during the Master's degree into practice and serves as a first professional experience in the world of linguistic services. After three months, I learned what it means to be a translator in a company, the different steps of the translation process, how to use the computer-assisted tools for translating used by professionals and research techniques. The importance of using terminology databases in technical translation became apparent in the 45 projects worked on throughout the course of the internship, each one with its specific linguistic features and terminology of their respective field. All the tools for translating used during the internship functioned with terminology databases and translation memories and through them it was possible to improve the efficiency and accuracy of the intern's translation skills. The use of CAT-Tools is nowadays absolutely essential for any translator that dedicates themselves to technical translation.

Protein synthesis is a fundamental process in gene expression. It involves four basic steps – initiation, elongation, termination, and ribosome recycling in bacteria. However, the step of initiation is the most regulated and rate-limiting step. Regulation at this step can save cellular energy and avoid toxicity due to the accretion of defective proteins. Bacteria have evolved with several mechanisms to ensure the fidelity of translation initiation. One of these mechanisms is the selection of the initiator tRNA (i-tRNA) at the P-site of the ribosome. The i-tRNA is endowed with unique structural features which facilitate its selection at the P-site. Two of the special properties of the i-tRNAs are the formylation of amino acid attached to it and, the highly conserved occurrence of the three consecutive G-C (3GC) base pairs in its anticodon stem. While the property of formylation facilitates its binding to the 30S ribosome, the 3GC pairs are important for its recognition in the ribosomal P-site and the maintenance of the fidelity of translation initiation. However, the 3GC base pairs mediated mechanisms used by the cell and the ribosome to ensure the fidelity of initiation are not fully understood. The mutations in the 3GC base pairs avert its selection at the P-site and its involvement in initiation. We made use of this phenotype in isolating E. coli mutants which allow translation initiation with the mutant 3GC i-tRNA. The characterization of many of the mutation(s) in these isolates has revealed the factors or pathways used by the cell to ensure the selection of canonical i-tRNA for initiation. Previously, the characterization of some of these suppressors unveiled that not only the ribosomal or the translation factors but also that cellular metabolism plays a vital role in ensuring the fidelity of translation initiation. For instance, a relatively high level of i-tRNA, modifications in 16S rRNA, optimum SD-aSD interaction, and rapid formylation of i-tRNA are critical factors to ensure the selection of i-tRNA at the P-site for initiation. The present set of studies involve the characterization of two more of the E. coli suppressor strains and the revelation of new factors that ensure the fidelity of translation initiation. One of the studies shows the role of N10-fTHF levels in the cell in the formylation of i-tRNA and suggests how the rapid speed of formylated i-tRNA avoids initiation with the mutant 3GC i-tRNA. Another study shows the role of large ribosomal subunit pseudouridine synthase D (RluD) in the fidelity of translation initiation. A point mutation in the C-terminal tail domain of RluD allows initiation with the 3GC mutant i-tRNA. The study suggests the possible role of the C-terminal tail of RluD in the biogenesis of the 30S subunit. Yet another study reveals the role of the -subunit of RNA polymerase (RpoA) in ensuring the initiation with i-tRNA. The point mutation in the C-terminal tail of RpoA allows translation initiation with the 3GC mutant i-tRNA. This study suggests the possibility of the impact of transcription-translation coupling in the fidelity of translation initiation.
Indian Institute of Science

The presence of the three consecutive GC base pairs in the anticodon stems of the initiator tRNAs (3GC base pairs) are a highly conserved and a vital feature of the initiator tRNAs in all the domains of life. How this feature is involved in the process of translation initiation is not very clear. The work described in this thesis entitled as “Initiation of protein synthesis: Role of the three consecutive GC base pairs in the anticodon stem of initiator tRNAs” provides new insights into the role of the 3GC base pairs in translation. Chapter 1 provides the relevant information regarding protein synthesis, initiator tRNA selection and the ribosome biogenesis. The experimental procedures as well as the materials used to carry out the study are mentioned in the Chapter 2. The remaining 3 chapters (Chapters 3 to 5) describe the results of the research findings on the role of 3GC base pairs in translation initiation. The brief summary of the research work described in thesis is summarized as follows. 1: Crosstalk between the initiator tRNA and the mRNA. Initiator tRNAs are special in their direct binding to the ribosomal P-site due to the hallmark occurrence of the three consecutive GC base pairs (3GC base pairs) in their anticodon stems. How the 3GC base pairs function in this role, has remained unsolved. The research described in Chapter 3 shows that mutations in either the mRNA or rRNA leading to extended interaction between the Shine-Dalgarno (SD) and anti-SD sequences compensate for the vital need of the 3GC base pairs in tRNAfMet for its function in Escherichia coli. In vivo, the 3GC mutant tRNAfMet occurred less abundantly in 70S ribosomes but normally on 30S subunits. However, the extended SD: anti-SD interaction increased its occurrence in 70S ribosomes. We propose that the 3GC base pairs play a critical role in tRNAfMet retention in ribosome during the conformational changes that mark the transition of 30S pre-initiation complex into elongation competent 70S complex. Furthermore, treating cells with kasugamycin, decreasing ribosome recycling factor (RRF) activity or increasing initiation factor 2 (IF2) levels enhanced initiation with the 3GC mutant transmits, suggesting that the 70S mode of initiation is less dependent on the 3GC base pairs in tRNAfMet. 2: Essentiality of the 3GC base pairs in the initiator tRNA. Translation initiation involves binding of mRNA to the small ribosomal subunit along with the initiation factors and the initiator tRNA followed by recruitment of 50S subunit to form 70S initiation complex. In eubacteria selection of initiator tRNA in the P- site of the ribosome mainly depends upon the formylation of the amino acid attached to it and the presence of 3GC base pairs in the anticodon stem of the tRNA. Unlike the property of formylation, which is conserved only in eubacteria, the presence of the 3GC base pairs in the initiator tRNA is highly conserved in all the three domains of life. The presence of natural variants lacking 1st or 3rd GC pair in mycoplasmal and rhizobial species implicates the essentiality of middle GC pair. This poses a question of why 3GC base pairs are retained in all three domains of life. In this study (Chapter 4), we show that the middle GC pair, particularly mid G (G30) residue is indeed essential for the initiator tRNA to sustain E. coli. However, the flanking GC pairs also contribute to the efficiency of the initiator tRNA function. Further, most of the mutants in the 3GC base pairs were abundant in 30S ribosomes but showed defect in the 70S ribosomal occupancy to varied extent suggesting their role in transition from 30S initiation complex (IC) to the 70S complex. Our in vivo genetic analysis shows that formylation of the amino acid attached to the initiator tRNA is mainly needed for the initial binding of the tRNA to the 30S ribosome while later steps are mainly dependent on the 3GC base pairs to form functional 70S complex. One such step could well be the eviction of IF3 from the 70S IC to convert it into an elongation competent 70S complex. Thus, the two highly conserved features of the eubacterial initiator tRNAs function at two distinct stages of initiation and help it to proceed through the various checkpoints during the step of initiation. Further, we provide evidence to suggest that the rhizobial and mycoplasmal species could do away with the full complement of 3 GC base pairs as they might have co-evolved ribosomes to utilize variant of initiator tRNA. 3: Role of the 3GC base pairs in the ribosome biogenesis. Ribosomes are the largest molecules of the cell which play a central role in the synthesis of the polypeptides and are also a hub of various regulatory mechanisms. The biogenesis of the ribosome is a complex and multistep process which requires proper folding of rRNAs as well as incorporation of more than 50 proteins. Any hampering in these processes can generate defective ribosomes which can be detrimental to the cell as they could lead to inefficient and/or mistranslation of mRNAs. As ribosomes are complex molecules, how the cell ensures their proper synthesis and their functional fidelity has remained unclear. In the genetic study described in Chapter 5, we find that the 3GC base pairs of the initiator tRNA play an important role in final maturation of ribosomes. Using a number of diverse genetic analyses in E. coli, we show that the formation of initiation complex by initiator tRNA contributes to ensure a proper biogenesis of the ribosomes. Ribosomal complexes defective in the formation of initiation complex were found to retain immature 5‟ and 3‟ ends of the 16S rRNAs in the complex.

Indiana University-Purdue University Indianapolis (IUPUI)
The DNA Base Excision Repair (BER) pathway repairs DNA damaged by endogenous and exogenous agents including chemotherapeutic agents. Removal of the damaged base by a DNA glycosylase creates an apurinic / apyrimidinic (AP) site. AP endonuclease1 (Ape1), a critical component in this pathway, hydrolyzes the phosphodiester backbone 5’ to the AP site to facilitate repair. Additionally, Ape1 also functions as a redox factor, known as Ref-1, to reduce and activate key transcription factors such as AP-1 (Fos/Jun), p53, HIF-1α and others. Elevated Ape1 levels in cancers are indicators of poor prognosis and chemotherapeutic resistance, and removal of Ape1 via methodology such as siRNA sensitizes cancer cell lines to chemotherapeutic agents. However, since Ape1 is a multifunctional protein, removing it from cells not only inhibits its DNA repair activity but also impairs its other functions. Our hypothesis is that a small molecule inhibitor of the DNA repair activity of Ape1 will help elucidate the importance (role) of its repair function in cancer progression as wells as tumor drug response and will also give us a pharmacological tool to enhance cancer cells’ sensitivity to chemotherapy. In order to discover an inhibitor of Ape1’s DNA repair function, a fluorescence-based high-throughput screening (HTS) assay was used to screen a library of drug-like compounds. Four distinct compounds (AR01, 02, 03 and 06) that inhibited Ape1’s DNA repair activity were identified. All four compounds inhibited the DNA repair activity of purified Ape1 protein and also inhibited Ape1’s activity in cellular extracts. Based on these and other in vitro studies, AR03 was utilized in cell culture-based assays to test our hypothesis that inhibition of the DNA repair activity of Ape1 would sensitize cancer cells to chemotherapeutic agents. The SF767 glioblastoma cell line was used in our assays as the chemotherapeutic agents used to treat gliobastomas induce lesions repaired by the BER pathway. AR03 is cytotoxic to SF767 glioblastoma cancer cells as a single agent and enhances the cytotoxicity of alkylating agents, which is consistent with Ape1’s inability to process the AP sites generated. I have identified a compound, which inhibits Ape1’s DNA repair activity and may have the potential in improving chemotherapeutic efficacy of selected chemotherapeutic agents as well as to help us understand better the role of Ape1’s repair function as opposed to its other functions in the cell.

La phosphorylation des protéines constitue l’une des plus importantes modifications post-traductionnelles (PTMs) et intervient dans de multiples processus physiologiques tels, la croissance, la différenciation cellulaire, l’apoptose, etc. En dépit de son importance, l’analyse des phosphoprotéines demeure une tâche difficile en raison de leur nature dynamique (car la phosphorylation des protéines est un processus réversible) et de leur faible abondance relative. En effet, la détermination des sites de phosphorylation est souvent difficile car les phosphopeptides sont souvent difficiles à détecter par des méthodes d’analyse chromatographique classique et par spectrométrie de masse (MS). De récentes études ont démontré que les nombreuses méthodes d’enrichissement de phosphopeptides existantes ne sont pas complètes, et que le nombre total de phosphopeptides détectés ne chevauchent pas complètement ces méthodes. C’est pour cela qu’il existe une nécessité de combler les lacunes des méthodes d’enrichissement existantes afin d’avoir des analyses phosphoprotéomiques plus complètes. Dans cette étude, nous avons utilisé les liquides ioniques (LI), plus particulièrement les sels d’imidazolium, comme une technique d’enrichissement alternative, dans le but de favoriser une extraction sélective de phosphopeptides présents en solution. Les sels d’imidazolium ont donc été utilisés en raison de leurs propriétés physico-chimiques "facilement" ajustables selon la nature des substituants sur le noyau imidazolium et la nature de l’anion. Les sels de monoimidazolium et de bis-imidazolium possédant respectivement des chaînes linéaires à 4, 12 et 16 atomes de carbone et ayant différents anions ont été synthétisés et utilisés pour effectuer des extractions liquide-liquide et solide-liquide des phosphopeptides en solution. Dans un premier temps, des extractions liquide-liquide ont été réalisées en utilisant un liquide ionique (LI) ayant une chaine linéaire de 4 atomes de carbone. Ces extractions réalisées avec le bis(trifluoromethanesulfonyl) amide de 3-butyl-1-methylimidazolium (BMIM-NTf2) et l’hexafluorophosphate de 3-butyl-1-methylimidazolium (BMIM-PF6) n’ont pas montré une extraction notable du PPS comparativement au PN. Dans un deuxième temps, des extractions solide-liquide ont été réalisées en fonctionnalisant des particules solides avec des sels d’imidazolium possédant des chaines linéaires de 12 ou 16 atomes de carbone. Ces extractions ont été faites en utilisant un phosphopentapeptide Ac-Ile-pTyr-Gly-Glu-Phe-NH2 (PPS) en présence de 2 analogues acides non-phosphorylés. Il a été démontré que les sels d’imidazolium à chaine C12 étaient meilleurs pour extraire le PPS que les deux autres peptides PN (Ac-Ile-Tyr-Gly-Glu-Phe-NH2) et PE (Ac-Glu-Tyr-Gly-Glu-Phe-NH2) L’électrophorèse capillaire (CE) et la chromatographie liquide à haute performance couplée à la spectrométrie de masse (LC-MS) ont été utilisées pour quantifier le mélange des trois peptides avant et après extraction ; dans le but de mesurer la sélectivité et l’efficacité d’extraction de ces peptides par rapport à la composition chimique du liquide ionique utilisé.
Protein phosphorylation is one of the most important post-translational modifications because it is involved in multiple physiological processes such as growth, differentiation, apoptosis, etc. Despite its importance, the analysis of phosphoproteins remains a difficult task due to their dynamic nature (phosphorylation of proteins is a reversible process) and their low abundance. Indeed, the determination of phosphorylation sites is difficult because phosphopeptides are often difficult to detect by conventional chromatographic analysis and by mass spectrometric (MS) methods. Recent studies have shown that the existing methods of enrichment of phosphopeptides are not complete, and the total number of phosphopeptides detected does not overlap completely with those detected by these methods. The gaps in existing enrichment methods need to be filled in order to have more complete phosphoproteomic analyses. In the current study, ionic liquids (IL), specifically imidazolium salts, have been used in an alternative enrichment technique with potential for selective extraction of phosphopeptides from solution. Imidazolium salts were chosen because their physicochemical properties are readily adjustable depending on the nature of the substituent attached to the imidazolium core and the counter-anion. Monoimidazolium and bis-imidazolium salts with linear chains having respectively 4, 12, and 16 carbon atoms and with different anions were synthesized and used to carry out liquid-liquid and solid-liquid extractions of a phosphorylated peptide from a solution. At first, liquid-liquid extractions were carried out using an ionic liquid (IL) with a linear chain of 4 carbon atoms. These extractions performed with bis (trifluoromethanesulfonyl) amide 3-butyl-1-methylimidazolium (BMIM-NTf2) and hexafluorophosphate 3-butyl-1-methylimidazolium (BMIM-PF6) did not show a considerable extraction of PPS comparatively to the PN. Secondly, solid-liquid extractions were done by first functionalizing solid-phase particles with the imidazolium salts. The extractions were carried out using the phosphopentapeptide Ac-pTyr-Ile-Gly-Glu-Phe-NH2 (PPS) and its acidic non-phosphorylated analogues. It has been shown that the C12 chain imidazolium salts were better to extract PPS than the other two peptides PN (Ac-Ile-Tyr-Gly-Glu-Phe-NH2) and PE (Ac-Glu-Tyr-Gly-Glu-Phe-NH2). The extraction efficiency of these peptides was estimated by capillary electrophoresis (CE) and high performance liquid chromatography coupled with mass spectrometry (LC-MS).

La traduction statistique vise l’automatisation de la traduction par le biais de modèles statistiques. Dans ce travail, nous relevons un des grands défis du domaine : la recherche (Brown et al., 1993). Les systèmes de traduction statistique de référence, tel Moses (Koehn et al., 2007), effectuent généralement la recherche en explorant l’espace des préfixes par programmation dynamique, une solution coûteuse sur le plan computationnel pour ce problème potentiellement NP-complet (Knight, 1999). Nous postulons qu’une approche par recherche locale (Langlais et al., 2007) peut mener à des solutions tout aussi intéressantes en un temps et un espace mémoire beaucoup moins importants (Russell et Norvig, 2010). De plus, ce type de recherche facilite l’incorporation de modèles globaux qui nécessitent des traductions complètes et permet d’effectuer des modifications sur ces dernières de manière non-continue, deux tâches ardues lors de l’exploration de l’espace des préfixes. Nos expériences nous révèlent que la recherche locale en traduction statistique est une approche viable, s’inscrivant dans l’état de l’art.
Statistical machine translation is a concerted effort towards the automation of the translation process. In the work presented here, we explore one of the major challenges of statistical machine translation: the search step (Brown et al., 1993). State of the art systems such as Moses (Koehn et al., 2007) search by exploring the prefix search space, a computationally costly solution to this potentially NP-complete problem (Knight, 1999). We propose that a local search approach can yield solutions which are qualitatively just as interesting, while keeping memory space and execution time at lower levels (Russell et Norvig, 2010). Furthermore, this type of search facilitates the use of global models for which a complete translation is needed and allows for non-continuous modifications, two tasks made difficult by exploring the prefix search space. The experiments we have conducted reveal that the use of local search during the search step in statistical machine translation is a viable, state of the art approach.