O propósito principal desta pesquisa é analisar, pelo viés da Análise de Discurso (AD), as condições de produção de um discurso sobre o sujeito aluno na Educação a Distância (EaD), a partir do modo como a legislação, o discurso acadêmico, via fórum de discussões, e a própria percepção do aluno o constituem como sujeito da EaD. Procuramos examinar as materialidades discursivas sobre a EaD: recortes de textos jurídicos, textos ministeriais e sequências discursivas dos sujeitos do campo da prática para entender as condições de produção dos discursos que institucionalizam e determinam um perfil para o aluno da EaD e, sobretudo, pelo questionamento sobre como os processos discursivos dessa institucionalização produzem mecanismos de identificação/apropriação ou tensionamento por parte dos sujeitos da prática em suas posições de sujeito. Voltamo-nos para o campo teórico da AD com a perspectiva de que conceitos como Condições de Produção (CP), Memória Discursiva, Interdiscurso, Intradiscurso, Posição de Sujeito (PS), Formação Discursiva (FD) e Formação Ideológica (FI) paramentem nosso gesto analítico para responder aos objetivos específicos desta pesquisa: (a) identificar as pistas da constituição do sujeito no discurso legal; (b) compreender qual é o lugar do aluno no discurso acadêmico e, em que medida este discurso reproduz os saberes que determinam o modos jurídicos de constituição do aluno da EaD; (c) analisar a própria percepção do sujeito aluno que busca formação na modalidade educacional a distância. Nossa proposição de análise parte da hipótese de que a Formação Ideológica Tecnicista (FIT) permeia a constituição dos saberes jurídicos da Formação Discursiva Jurídica da EaD (FDJEaD) e da Formação Discursiva Ministerial da EaD (FDMEaD) e de que esses saberes determinantes da forma-sujeito dessas FD’s são apropriados/reproduzidos ou tensionados pelos sujeitos do campo da prática, ao ocuparem posições de sujeito na Formação Discursiva Institucional da EaD (FDIEaD). A noção de reprodução é definida pelo conceito de paráfrase discursiva compreendida nos termos do que Pêcheux (1997b) elencou como a determinação do interdiscurso no intradiscurso. Procuramos mostrar, por conseguinte, que as sequências discursivas (Sd) produzidas por alunos, acadêmicos e tutores que ocupam posições-sujeito na FDIEaD podem evidenciar processos de identificação, ou contra-identificação em relação aos saberes institucionalizados nos discursos oficiais. Motivamos, com essa análise, um olhar mais atento para o atravessamento do interdiscurso (memória discursiva) no processo de apropriação de saberes ideológicos que ressaltam na materialidade da língua – intradiscurso. Em todos os recortes (rec.) e sequências discursivas (sdr) analisados percebemos claramente o processo de intervenção da ideologia tecnicista na constituição de saberes das três FD’s analisadas (FDJEaD, FDMEaD, FDIEaD) e, sobremaneira, detectamos mecanismos linguísticos de identificação e contra-identificação das posições-sujeito das práticas de ensino-aprendizagem (acadêmicos, alunos, tutores) em relação ao discurso superestrutural (jurídico e ministerial). / The main purpose of this research is to analyse, from the perspective of Discourse Analysis (DA), the condition of the production of a discourse about the student talking into Distance Learning, account how the legislation, the academic discourse (through forum) and the studenr self-perception turn him into the subject of the Distance Learning (DL). We seek to examine the discursive materialities about the DL (excerpts of legal texts, ministerial texts and discursive sequences of subjects from the field of practice) to understand the conditions of production of the discourses that institutionalize and determine a profile to the DL’s student and above all by the questioning of how the discursive processes of this institutionalization produce mechanisms of identification/appropriation or tensioning from the subjects of the practice in their positions as subjects.We turn to the theoretical field of the DA with the perspective that concepts such as Production Conditions (PC), Discursive Memory, Interdiscourse, Intradiscourse, Subject Position (SP), Discursive Training (DT) and Ideological Training (IT) will embody our analytical gesture to answer the specific goals of this research: (a) to identify the clues of the constitution of the subject in the legal discourse; (b) to understand the place of the student in the academic discourse and in which way this discourse reproduces the knowledge that defines the legal ways of constitution of the DLstudent; (c) to analyse the self-perception of the student that searches education in the distance learning modality. Our analysis proposition starts with the hypothesis that the Technicist Ideological Training (TIT) permeates the constitution of the legal knowledge of the DL’s Legal Discursive Training (LDTDL) and the DL’s Ministerial Discursive Training (MDTDL) and that these determining knowledge of the form-subject of these Discursive Training is appropriated/reproduced or tensioned by the subjects of the field of practice once they occupy the subject position in the Institutional Discursive Training of the DL (IDTDL). The notion of the reproduction is defined by the concept of discursive paraphrase understood in the terms that Pêcheux (1997b) listed out as a determination of the interdiscourse in the intradiscourse. Thus, we seek to show that the discursive sequences (DS) produced by the students, scholars and tutors that occupy subject-positions in the IDTDL can suggest the identification processes or counter-identification processes in relation to the institutionalized knowledges in the official discourses. The analysis points to a closer look to the crossing speech (discursive memory) in the process of the appropriation of ideological knowledge that highlights the language materiality – intradiscourse. In every excerpt and discursive sequences (SD) analysed we clearly perceived the process of the intervention of the technicist ideology in the constitution of the knowledgeof the three DTs that were analysed (LDTDL, MDTDL, IDTDL) and we detected linguistic mechanisms of identification and counteridentification from the subject-positions of the teaching and learning practices (scholars, students, tutors) relative to the super structural discourse (legal and ministerial).

O propósito principal desta pesquisa é analisar, pelo viés da Análise de Discurso (AD), as condições de produção de um discurso sobre o sujeito aluno na Educação a Distância (EaD), a partir do modo como a legislação, o discurso acadêmico, via fórum de discussões, e a própria percepção do aluno o constituem como sujeito da EaD. Procuramos examinar as materialidades discursivas sobre a EaD: recortes de textos jurídicos, textos ministeriais e sequências discursivas dos sujeitos do campo da prática para entender as condições de produção dos discursos que institucionalizam e determinam um perfil para o aluno da EaD e, sobretudo, pelo questionamento sobre como os processos discursivos dessa institucionalização produzem mecanismos de identificação/apropriação ou tensionamento por parte dos sujeitos da prática em suas posições de sujeito. Voltamo-nos para o campo teórico da AD com a perspectiva de que conceitos como Condições de Produção (CP), Memória Discursiva, Interdiscurso, Intradiscurso, Posição de Sujeito (PS), Formação Discursiva (FD) e Formação Ideológica (FI) paramentem nosso gesto analítico para responder aos objetivos específicos desta pesquisa: (a) identificar as pistas da constituição do sujeito no discurso legal; (b) compreender qual é o lugar do aluno no discurso acadêmico e, em que medida este discurso reproduz os saberes que determinam o modos jurídicos de constituição do aluno da EaD; (c) analisar a própria percepção do sujeito aluno que busca formação na modalidade educacional a distância. Nossa proposição de análise parte da hipótese de que a Formação Ideológica Tecnicista (FIT) permeia a constituição dos saberes jurídicos da Formação Discursiva Jurídica da EaD (FDJEaD) e da Formação Discursiva Ministerial da EaD (FDMEaD) e de que esses saberes determinantes da forma-sujeito dessas FD’s são apropriados/reproduzidos ou tensionados pelos sujeitos do campo da prática, ao ocuparem posições de sujeito na Formação Discursiva Institucional da EaD (FDIEaD). A noção de reprodução é definida pelo conceito de paráfrase discursiva compreendida nos termos do que Pêcheux (1997b) elencou como a determinação do interdiscurso no intradiscurso. Procuramos mostrar, por conseguinte, que as sequências discursivas (Sd) produzidas por alunos, acadêmicos e tutores que ocupam posições-sujeito na FDIEaD podem evidenciar processos de identificação, ou contra-identificação em relação aos saberes institucionalizados nos discursos oficiais. Motivamos, com essa análise, um olhar mais atento para o atravessamento do interdiscurso (memória discursiva) no processo de apropriação de saberes ideológicos que ressaltam na materialidade da língua – intradiscurso. Em todos os recortes (rec.) e sequências discursivas (sdr) analisados percebemos claramente o processo de intervenção da ideologia tecnicista na constituição de saberes das três FD’s analisadas (FDJEaD, FDMEaD, FDIEaD) e, sobremaneira, detectamos mecanismos linguísticos de identificação e contra-identificação das posições-sujeito das práticas de ensino-aprendizagem (acadêmicos, alunos, tutores) em relação ao discurso superestrutural (jurídico e ministerial). / The main purpose of this research is to analyse, from the perspective of Discourse Analysis (DA), the condition of the production of a discourse about the student talking into Distance Learning, account how the legislation, the academic discourse (through forum) and the studenr self-perception turn him into the subject of the Distance Learning (DL). We seek to examine the discursive materialities about the DL (excerpts of legal texts, ministerial texts and discursive sequences of subjects from the field of practice) to understand the conditions of production of the discourses that institutionalize and determine a profile to the DL’s student and above all by the questioning of how the discursive processes of this institutionalization produce mechanisms of identification/appropriation or tensioning from the subjects of the practice in their positions as subjects.We turn to the theoretical field of the DA with the perspective that concepts such as Production Conditions (PC), Discursive Memory, Interdiscourse, Intradiscourse, Subject Position (SP), Discursive Training (DT) and Ideological Training (IT) will embody our analytical gesture to answer the specific goals of this research: (a) to identify the clues of the constitution of the subject in the legal discourse; (b) to understand the place of the student in the academic discourse and in which way this discourse reproduces the knowledge that defines the legal ways of constitution of the DLstudent; (c) to analyse the self-perception of the student that searches education in the distance learning modality. Our analysis proposition starts with the hypothesis that the Technicist Ideological Training (TIT) permeates the constitution of the legal knowledge of the DL’s Legal Discursive Training (LDTDL) and the DL’s Ministerial Discursive Training (MDTDL) and that these determining knowledge of the form-subject of these Discursive Training is appropriated/reproduced or tensioned by the subjects of the field of practice once they occupy the subject position in the Institutional Discursive Training of the DL (IDTDL). The notion of the reproduction is defined by the concept of discursive paraphrase understood in the terms that Pêcheux (1997b) listed out as a determination of the interdiscourse in the intradiscourse. Thus, we seek to show that the discursive sequences (DS) produced by the students, scholars and tutors that occupy subject-positions in the IDTDL can suggest the identification processes or counter-identification processes in relation to the institutionalized knowledges in the official discourses. The analysis points to a closer look to the crossing speech (discursive memory) in the process of the appropriation of ideological knowledge that highlights the language materiality – intradiscourse. In every excerpt and discursive sequences (SD) analysed we clearly perceived the process of the intervention of the technicist ideology in the constitution of the knowledgeof the three DTs that were analysed (LDTDL, MDTDL, IDTDL) and we detected linguistic mechanisms of identification and counteridentification from the subject-positions of the teaching and learning practices (scholars, students, tutors) relative to the super structural discourse (legal and ministerial).

This work aims to produce a knowledge about French Discourse Analysis (DA) through Eni P. Orlandi s theoretical production and, as specific goal, to form a possibility of interpretation for the notions of paraphrase and metaphor, from A linguagem e seu funcionamento: as formas do discurso (1983) and establishing relations with Interpretação: autoria, leitura e efeitos do trabalho simbólico (1996) e Análise de Discurso: princípios e procedimentos (1999), with the purpose of reflecting about the development of DA in Brazil, by the analysis of principles and procedures able to justify its disciplinary consistency. We suppose that metalinguistics, as a representation of a domain of knowledge, may be related to epistemology and to the thought about science. However, the representation made by metalinguistics differs from the positive epistemology of science which admits the transparency of epistemic modality, the objectivity and universality of knowledge. Metalinguistics historically built supports the paradox of knowledge and dissimulates the coherence of scientific knowledge which no longer refers itself to the truth, but to scientific communities. These communities point to exclusion, the distribution of knowledge; and inclusion, trough the recognition of equals who can validate its results. The communities also justify the discipline which functions as a possibility to legitimate its existence. The research permits to elaborate a thesis around the idea of paraphrase and metaphor constitute itself, from the disciplinary point of view, as a procedure of appropriation which moves itself rom content analysis , associated to Social Sciences; and moves itself from the signification of Structural Linguistics; making them to work as a discourse, not as a text. The theoretical movement of this notions paraphrase and metaphor - is justified if science is thought as apparently coherent, and justify itself if it s thought in terms of history, thus these procedures place paraphrase and metaphor in the tradition of studies of Language in Brazil. / Este trabalho tem como objetivo geral compor um conhecimento sobre a Análise de Discurso (AD) de orientação francesa pela produção teórica de Eni P. Orlandi e, como objetivo específico, elaborar uma possibilidade de leitura das noções de paráfrase e de polissemia, a partir de A linguagem e seu funcionamento: as formas do discurso (1983) e estabelecendo relações com Interpretação: autoria, leitura e efeitos do trabalho simbólico (1996) e Análise de Discurso: princípios e procedimentos (1999), com o propósito de refletir sobre o desenvolvimento da AD no Brasil, ao identificar princípios e procedimentos responsáveis por sua consistência disciplinar. Parte-se do pressuposto de que a metalinguagem enquanto representação de um domínio do saber pode ser relacionada à epistemologia e ao pensamento elaborado sobre a ciência de uma maneira geral. Mas a representação operada por ela diferencia-se da epistemologia positiva da ciência que admite a transparência da modalidade epistêmica, a objetividade, a universalidade e a intangibilidade do conhecimento. A metalinguagem historicamente construída suporta a natureza paradoxal do conhecimento e dissimula a coerência do conhecimento científico que não se refere mais à verdade, mas a comunidades em torno das quais se organiza. A existência de comunidades do conhecimento indiciam, portanto, a exclusão, a demarcação de fronteiras dos saberes; e a inclusão, pelo reconhecimento dos pares que podem validar seus resultados. Justificam também a existência do disciplinar que funciona como uma possibilidade de legitimar a existência de uma tal comunidade. A trajetória realizada no presente estudo nos permite elaborar a tese de que as noções de paráfrase e de polissemia, constituem-se, a partir do ponto de vista disciplinar, em um procedimento de reapropriação/fundação que as desloca do conteudismo, associado às Ciências Sociais; e da significação associada à Lingüística saussureana; fazendo-as funcionar discursivamente. O movimento teórico que vai da noção de efeito metafórico à paráfrase e polissemia justifica-se se a ciência for pensada em termos de coerência aparente, não de proposições verdadeiras; e justifica-se também se a ciência for pensada em sua historicidade, assim, essa operação localiza a paráfrase e a polissemia na tradição dos estudos da linguagem no caso específico do Brasil.

Sentiment analysis is a field of computational linguistics involving identification, extraction, and classification of opinions, sentiments, and emotions expressed in natural language. Sentiment classification algorithms aim to identify whether the author of a text has a positive or a negative opinion about a topic. One of the main indicators which help to detect the opinion are the words used in the texts. Needless to say, the sentiments expressed in the texts also depend on the syntactic structure and the discourse context. Supervised machine learning approaches to sentiment classification were shown to achieve good results. Classifying texts by emotions requires finer-grained analysis than sentiment classification. In this thesis, we explore the task of emotion and mood classification for blog postings. We propose a novel approach that uses the hierarchy of possible moods to achieve better results than a standard flat classification approach. We also show that using sentiment orientation features improves the performance of classification. We used the LiveJournal blog corpus as a dataset to train and evaluate our method. Another contribution of this work is extracting paraphrases for emotion terms based on the six basics emotions proposed by Ekman (\textit{happiness, anger, sadness, disgust, surprise, fear}). Paraphrases are different ways to express the same information. Algorithms to extract and automatically identify paraphrases are of interest from both linguistic and practical points of view. Our paraphrase extraction method is based on a bootstrapping algorithms that starts with seed words. Unlike in previous work, our algorithm does not need a parallel corpus. In Natural Language Generation (NLG), paraphrasing is employed to create more varied and natural text. In our research, we extract paraphrases for emotions, with the goal of using them to automatically generate emotional texts (such as friendly or hostile texts) for conversations between intelligent agents and characters in educational games. Nowadays, online services are popular in many disciplines such as: e-learning, interactive games, educational games, stock market, chat rooms and so on. NLG methods can be used in order to generate more interesting and normal texts for such applications. Generating text with emotions is one of the contributions of our work. In the last part of this thesis, we give an overview of NLG from an applied system's points of view. We discuss when NLG techniques can be used; we explained the requirements analysis and specification of NLG systems. We also, describe the main NLG tasks of content determination, discourse planning, sentence aggregation, lexicalization, referring expression generation, and linguistic realisation. Moreover, we describe our Authoring Tool that we developed in order to allow writers without programming skills to automatically generate texts for educational games. We develop an NLG system that can generate text with different emotions. To do this, we introduce our pattern-based model for generation. We show our model starts with initial patterns, then constructs extended patterns from which we choose ``final'' patterns that are suitable for generating emotion sentences. A user can generate sentences to express the desired emotions by using our patterns. Alternatively, the user can use our Authoring Tool to generate sentences with emotions. Our acquired paraphrases will be employed by the tool in order to generate more varied outputs.

This study focuses on a comparative analysis of two piano paraphrases, Pastiche on Habanera from ‘Carmen’ by Kaikhosru Sorabji and Variations on a Theme from Bizet’s ‘Carmen’ by Vladimir Horowitz. These compositions idiomatically distort the original material in a manner that was not explored up to the moment of their respective conception. They expose each composer’s free compositional approach, reflecting musical freedom rooted in the originality of their musical thinking. The aesthetic uniqueness of these two compositions strongly stimulates and justifies academic interest to explore their technical construction, musical differences, and artistic significance. This study proposes to undertake a comparative study of these two compositions, analyzing (1) aspects of the musical character, which are linked with embellishment, or rearrangement of original material, and (2) differences in performance approach based on recorded examples and critical observations by others of the performances of these works by Sorabji and Horowitz.

No description available.

Undoubtedly, Totentanz has been one of the most famous works by Franz Liszt. Totentanz has been recorded by many pianists and addressed in much of the vast literature about Liszt and his works; however, little research has been focused on this work. Most studies of Totentanz address only the historical background of the piece in relation to the theme based on Dies irae. Currently, there are no specific studies about the solo piano or two piano versions and only one recording was located. Liszt's own piano solo transcription of this famous work is an excellent addition to the concert repertoire. Totentanz consists of six variations that include canonic and fugato sections. The main theme is based on the Gregorian chant Dies irae, a melody that has been used by many other composers, most notably Berlioz in Witches Sabbath of Symphonie fantastique, op. 14 and Rachmaninoff in Rhapsody on a Theme of Paganini. This study contains five chapters. Chapters I and II provide background information, historical background and influences of Totentanz. Chapter III presents an outline of Liszt's achievement as a transcriber. Liszt revised his own works numerous times from the 1840s and 1850s, including Transcendental Etudes, Paganini Etudes, and piano and orchestra works. Like in the case of Totentanz, transcribed form piano and orchestra into piano solo, Liszt transcribed and paraphrased hundreds of other composers' works as well. Chapter IV discusses and compares the two main versions for solo piano and piano and orchestra. Form and harmonic language in particular the use of tritone in Totentanz is discussed. The adjustment required in transcribing the work for piano solo is discussed in detail, followed by a conclusion.

This thesis aims to contribute to the improvement of the paraphrasing competence, or the ability to produce synonymous sentences, of advanced learners of French as a second language (L2). While possessing solid grammatical skills, advanced learners still produce lexical and stylistic errors and experience difficulties in reformulating their discourse. Paraphrasing competence is crucial for a language learner since it helps him to present an idea more clearly or circumvent a lexical gap, while allowing the learner to express the same meaning in multiple ways. The theoretical framework adopted is Meaning-Text lin-guistic theory (MTT), that attaches a great deal of importance to the description of lexical and paraphrastic relations and uses well-developped formal tools for the modeling of these relations. The thesis has two objectives. The first consists in elaborating a methodology for evalua-ting the paraphrasing competence of a speaker. The notion of paraphrasing competence is defined, and a test assessing the ability to paraphrase is created and administered to An-glophone learners of French and native speakers. Quantitative and qualitative criteria to establish four levels of paraphrasing competence – elementary, intermediate, advanced and superior - are identified. Characteristics of advanced learners’ paraphrasing competence are described. While pro-ducing all types of paraphrases, the L2 learner uses less lexico-syntactic paraphrasing operations, such as synonymic, antonymic or derivative substitutions. His paraphrasing means are less diverse than those of a native speaker. He also produces a high number of paraphrases with lexical and syntactic errors. The second objective of the thesis is to propose a didactic method of teaching paraphrase, taking into consideration the challenges that paraphrasing poses for L2 learners, identified by means of the paraphrasing test. Concepts of the Meaning-Text Theory are presented in a user-friendly format so that the theoretical linguistic knowledge becomes more accessible to L2 learners. A structured and explicit teaching approach is adopted, based on: (i) a selection of key concepts pertaining to the paraphrase and their definitions ; (ii) a selection of MTT paraphrasing rules and their explanation; (iii) paraphrasing exercices.The advocated pedagogical approach may serve as a reference tool for paraphrase didactics in French.

Submitted by Luciana Sebin (lusebin@ufscar.br) on 2016-10-05T18:38:23Z No. of bitstreams: 1 DissPCP.pdf: 1921482 bytes, checksum: 5298cc1a066e0cfe217b2b9c61076e65 (MD5) / Approved for entry into archive by Marina Freitas (marinapf@ufscar.br) on 2016-10-14T14:13:08Z (GMT) No. of bitstreams: 1 DissPCP.pdf: 1921482 bytes, checksum: 5298cc1a066e0cfe217b2b9c61076e65 (MD5) / Approved for entry into archive by Marina Freitas (marinapf@ufscar.br) on 2016-10-14T14:13:18Z (GMT) No. of bitstreams: 1 DissPCP.pdf: 1921482 bytes, checksum: 5298cc1a066e0cfe217b2b9c61076e65 (MD5) / Made available in DSpace on 2016-10-14T14:13:28Z (GMT). No. of bitstreams: 1 DissPCP.pdf: 1921482 bytes, checksum: 5298cc1a066e0cfe217b2b9c61076e65 (MD5) Previous issue date: 2016-03-04 / Não recebi financiamento / Use different words to express/convey the same message is a necessity in any natural language and, as such, should be investigated in research in Natural Language Processing (NLP). When it is just a simple word, we say that the interchangeable words are synonyms; while the term paraphrase is used to express a more general idea and that also may involve more than one word. For example, the sentences "the light is red" and "the light is closed" are examples of paraphrases as "sign" and "traffic light" represent synonymous in this context. Proper treatment of paraphrasing is important in several NLP applications, such as Machine Translation, which paraphrases can be used to increase the coverage of Statistical Machine Translation systems; on Multidocument Summarization, where paraphrases identification allows the recognition of repeated information; and Natural Language Generation, where the generation of paraphrases allows creating more varied and fluent texts. The project described in this document is intended to verify that is possible to learn, in an incremental and automatic way, paraphrases in words level from a bilingual parallel corpus, using Never-Ending Machine Learning (NEML) strategy and the Internet as a source of knowledge. The NEML is a machine learning strategy, based on how humans learn: what is learned previously can be used to learn new information and perhaps more complex in the future. Thus, the NEML has been applied together with the strategy for paraphrases extraction proposed by Bannard and Callison-Burch (2005) where, from bilingual parallel corpus, paraphrases are extracted using a pivot language. In this context, it was developed NEPaL (Never-Ending Paraphrase Learner) AMSF system responsible for: (1) extract the internet texts, (2) align the text using a pivot language, (3) rank the candidates according to a classification model and (4) use the knowledge to produce a new classifier model and therefore gain more knowledge restarting the never-ending learning cycle. / Usar palavras diferentes para expressar/transmitir a mesma mensagem é uma necessidade em qualquer língua natural e, como tal, deve ser investigada nas pesquisas em Processamento de Língua Natural (PLN). Quando se trata apenas de uma palavra simples, dizemos que as palavras intercambiáveis são sinônimos; enquanto o termo paráfrase é utilizado para expressar uma ideia mais geral e que pode envolver também mais de uma palavra. Por exemplo, as sentenças “o sinal está vermelho” e “o semáforo está fechado” são exemplo de paráfrases enquanto “sinal” e “semáforo” representam sinônimos, nesse contexto. O tratamento adequado de paráfrases é importante em diversas aplicações de PLN, como na Tradução Automática, onde paráfrases podem ser utilizadas para aumentar a cobertura de sistemas de Tradução Automática Estatística; na Sumarização Multidocumento, onde a identificação de paráfrases permite o reconhecimento de informações repetidas; e na Geração de Língua Natural, onde a geração de paráfrases permite criar textos mais variados e fluentes. O projeto descrito neste documento visa verificar se é possível aprender, de modo incremental e automático, paráfrases em nível de palavras a partir de corpus paralelo bilíngue, utilizando a estratégia de Aprendizado de Máquina Sem-fim (AMSF) e a Internet como fonte de conhecimento. O AMSF é uma estratégia de Aprendizado de Máquina, baseada na forma como os humanos aprendem: o que é aprendido previamente pode ser utilizado para aprender informações novas e talvez mais complexas, futuramente. Para tanto, o AMSF foi aplicado juntamente com a estratégia para a extração de paráfrases proposta por Bannard e Callison-Burch (2005) onde, a partir de corpus paralelo bilíngue, paráfrases são extraídas utilizando um idioma pivô. Nesse contexto, foi desenvolvido o NEPaL (Never-Ending Paraphrase Learner), sistema de AMSF responsável por: (1) extrair textos da internet, (2) alinhar os textos utilizando um idioma pivô, (3) classificar as candidatas de acordo com um modelo de classificação e (4) utilizar o conhecimento para produzir um novo modelo classificador e, consequentemente, adquirir mais conhecimento reiniciando o ciclo de aprendizado sem-fim.

Mémoire numérisé par la Direction des bibliothèques de l'Université de Montréal.