Les thèmes principaux abordés dans cette thèse sont l'utilisation de méthodes d'apprentissage actif et d'apprentissage profond dans le contexte du traitement de documents multimodaux. Les contributions proposées dans cette thèse abordent ces deux thèmes. Un système d'apprentissage actif a été introduit pour permettre une annotation plus efficace des émissions de télévision grâce à la propagation des étiquettes, à l'utilisation de données multimodales et à des stratégies de sélection efficaces. Plusieurs scénarios et expériences ont été envisagés dans le cadre de l'identification des personnes dans les vidéos, en prenant en compte l'utilisation de différentes modalités (telles que les visages, les segments de la parole et le texte superposé) et différentes stratégies de sélection. Le système complet a été validé au cours d'un ``test à blanc'' impliquant des annotateurs humains réels.Une deuxième contribution majeure a été l'étude et l'utilisation de l'apprentissage profond (en particulier les réseaux de neurones convolutifs) pour la recherche d'information dans les vidéos. Une étude exhaustive a été réalisée en utilisant différentes architectures de réseaux neuronaux et différentes techniques d'apprentissage telles que le réglage fin (fine-tuning) ou des classificateurs plus classiques comme les SVMs. Une comparaison a été faite entre les caractéristiques apprises (la sortie des réseaux neuronaux) et les caractéristiques plus classiques (``engineered features''). Malgré la performance inférieure des seconds, une fusion de ces deux types de caractéristiques augmente la performance globale.Enfin, l'utilisation d'un réseau neuronal convolutif pour l'identification des locuteurs à l'aide de spectrogrammes a été explorée. Les résultats ont été comparés à ceux obtenus avec d'autres systèmes d'identification de locuteurs récents. Différentes approches de fusion ont également été testées. L'approche proposée a permis d'obtenir des résultats comparables à ceux certains des autres systèmes testés et a offert une augmentation de la performance lorsqu'elle est fusionnée avec la sortie du meilleur système. / The main topics of this thesis include the use of active learning-based methods and deep learning in the context of retrieval of multimodal documents. The contributions proposed during this thesis address both these topics. An active learning framework was introduced, which allows for a more efficient annotation of broadcast TV videos thanks to the propagation of labels, the use of multimodal data and selection strategies. Several different scenarios and experiments were considered in the context of person identification in videos, including using different modalities (such as faces, speech segments and overlaid text) and different selection strategies. The whole system was additionally validated in a dry run involving real human annotators.A second major contribution was the investigation and use of deep learning (in particular the convolutional neural network) for video retrieval. A comprehensive study was made using different neural network architectures and training techniques such as fine-tuning or using separate classifiers like SVM. A comparison was made between learned features (the output of neural networks) and engineered features. Despite the lower performance of the engineered features, fusion between these two types of features increases overall performance.Finally, the use of convolutional neural network for speaker identification using spectrograms is explored. The results are compared to other state-of-the-art speaker identification systems. Different fusion approaches are also tested. The proposed approach obtains comparable results to some of the other tested approaches and offers an increase in performance when fused with the output of the best system.

I fritidshemmen möts de flerspråkiga barnen av fritidspedagoger och svenskfödda barn. Enligt undersökningar vi tagit del av är det viktigt att bemötandet av de flerspråkiga barnen är genomtänkt och anpassat efter deras behov och kulturer. Det måste också ske med ett stimulerande arbetssätt. Undersökningarna visar även att språket är en förutsättning för att kunna lära sig nya saker, att kunna kommunicera med andra och att kunna stärka sin självkänsla och identitet. Syftet i vår undersökning är att belysa hur fritidspedagoger kan underlätta språkutveckling och integrering i fritidshemmet för flerspråkiga barn. Undersökningen tar reda på vilka arbetssätt som kan användas samt vilka möjligheter kontra vilka hinder som finns med flerspråkigheten i verksamheten. För att få reda på undersökningens alla frågor intervjuades 12 fritidspedagoger som varit verksamma i mer än 3 år. Denna undersökning är en kvalitativ studie. Resultatet visar att arbetssätten varierar hos fritidspedagogerna, men det finns också arbetssätt som de flesta av fritidspedagogerna har gemensamt. Undersökningen visar att det finns fler möjligheter än hinder med flerspråkiga barn i fritidshemmet.

Este trabalho investiga técnicas de extração de termos multipalavra a partir de corpora comparáveis, que são conjuntos de textos em duas (ou mais) línguas sobre o mesmo domínio. A extração de termos, especialmente termos multipalavra é muito importante para auxiliar a criação de terminologias, ontologias e o aperfeiçoamento de tradutores automáticos. Neste trabalho utilizamos um corpus comparável português/inglês e queremos encontrar termos e seus equivalentes em ambas as línguas. Para isso começamos com a extração dos termos separadamente em cada língua, utilizando padrões morfossintáticos para identificar os n-gramas (sequências de n palavras) mais prováveis de serem termos importantes para o domínio. A partir dos termos de cada língua, utilizamos o contexto, isto é, as palavras que ocorrem no entorno dos termos para comparar os termos das diferentes línguas e encontrar os equivalentes bilíngues. Tínhamos como objetivos principais neste trabalho fazer a identificação monolíngue de termos, aplicar as técnicas de alinhamento para o português e avaliar os diferentes parâmetros de tamanho e tipo (PoS utilizados) de janela para a extração de contexto. Esse é o primeiro trabalho a aplicar essa metodologia para o Português e apesar da falta de alguns recursos léxicos e computacionais (como dicionários bilíngues e parsers) para essa língua, conseguimos alcançar resultados comparáveis com o estado da arte para trabalhos em Francês/Inglês. / This work investigates techniques for multiword term extraction from comparable corpora, which are sets of texts in two (or more) languages about the same topic. Term extraction, specially multiword terms is very important to help the creation of terminologies, ontologies and the improvement of machine translation. In this work we use a comparable corpora Portuguese/ English and want to find terms and their equivalents in both languages. To do this we start with separate term extraction for each language. Using morphossintatic patterns to identify n-grams (sequences of n words) most likely to be important terms of the domain. From the terms of each language, we use their context, i. e., the words that occurr around the term to compare the terms of different languages and to find the bilingual equivalents. We had as main goals in this work identificate monolingual terms, apply alignment techniques for Portuguese and evaluate the different parameters of size and type (used PoS) of window to the context extraction. This is the first work to apply this methodology to Portuguese and in spite of the lack of lexical and computational resources (like bilingual dictionaries and parsers) for this language, we achieved results comparable to state of the art in French/English.

The purpose of this study is to examine the effects of bilingualism in short-term memory (STM) compared to monolingualism with individuals who have Down syndrome. Five tasks were used for STM skills comparison between monolingual and bilingual participants. Sixteen participants between the ages of 13 to 37 were included in this study. Participants were divided based on their language groups. The experimental tasks consisted of non-verbal activities to examine visual (RVDLT) and spatial (Corsi) STM; as well as three verbal STM tasks (RAVLT, WMS, and Digits). The results showed that bilinguals acquired higher overall correct responses, with a significant difference found in visual STM performance and a trend towards significance in verbal logical STM. Thus, the findings of this study support the bilingual advantage theory. This study will enhance understanding in memory capacity of bilingual individuals with Down syndrome for potential implications to put into practice in clinical intervention strategies.

abstract: This project investigated second language writing teachers’ writing assessment literacy by looking at teachers’ practices of electronic writing portfolios (e-WPs), as well as the sources that shape L2 writing teachers’ knowledge of e-WPs in the context of multilingual First-Year Composition (FYC) classrooms. By drawing on Borg’s (2003) theory of teacher cognition and Crusan, Plakans, and Gebril’s (2016) definition of assessment literacy, I define L2 teachers’ writing assessment literacy as teachers’ knowledge, beliefs and practices of a particular assessment tool, affected by institutional factors. While teachers are the main practitioners who help students create e-WPs (Hilzensauer & Buchberger, 2009), studies on how teachers actually incorporate e-WPs in classes and what sources may influence teachers’ knowledge of e-WPs, are scant. To fill in this gap, I analyzed data from sixteen teachers’ semi-structured interviews. Course syllabi were also collected to triangulate the interview data. The interview results indicated that 37.5 % of the teachers use departmental e-WPs with the goal of guiding students throughout their writing process. 43.7 % of the teachers do not actively use e-WPs and have students upload their writing projects only to meet the writing program’s requirement at the end of the semester. The remaining 18.7 % use an alternative platform other than the departmental e-WP platform, throughout the semester. Sources influencing teachers’ e-WP knowledge included teachers’ educational and work experience, technical difficulties in the e-WP platform, writing program policies and student reactions. The analysis of the course syllabi confirmed the interview results. Based on the findings, I argue that situated in the context of classroom assessment, institutional factors plus teachers’ insufficient knowledge of e-WPs limit the way teachers communicate with students, whose reactions cause teachers to resist e-WPs. Conversely, teachers’ sufficient knowledge of e-WPs enables them to balance the pressure from the institutional factors, generating positive reactions from the students. Students’ positive reactions encourage teachers to accept the departmental e-WPs or use similar alternative e-WP platforms. Pedagogical implications, limitations of the study and suggestions for future research are reported to conclude the dissertation. / Dissertation/Thesis / Doctoral Dissertation Linguistics and Applied Linguistics 2018

Este trabalho investiga técnicas de extração de termos multipalavra a partir de corpora comparáveis, que são conjuntos de textos em duas (ou mais) línguas sobre o mesmo domínio. A extração de termos, especialmente termos multipalavra é muito importante para auxiliar a criação de terminologias, ontologias e o aperfeiçoamento de tradutores automáticos. Neste trabalho utilizamos um corpus comparável português/inglês e queremos encontrar termos e seus equivalentes em ambas as línguas. Para isso começamos com a extração dos termos separadamente em cada língua, utilizando padrões morfossintáticos para identificar os n-gramas (sequências de n palavras) mais prováveis de serem termos importantes para o domínio. A partir dos termos de cada língua, utilizamos o contexto, isto é, as palavras que ocorrem no entorno dos termos para comparar os termos das diferentes línguas e encontrar os equivalentes bilíngues. Tínhamos como objetivos principais neste trabalho fazer a identificação monolíngue de termos, aplicar as técnicas de alinhamento para o português e avaliar os diferentes parâmetros de tamanho e tipo (PoS utilizados) de janela para a extração de contexto. Esse é o primeiro trabalho a aplicar essa metodologia para o Português e apesar da falta de alguns recursos léxicos e computacionais (como dicionários bilíngues e parsers) para essa língua, conseguimos alcançar resultados comparáveis com o estado da arte para trabalhos em Francês/Inglês. / This work investigates techniques for multiword term extraction from comparable corpora, which are sets of texts in two (or more) languages about the same topic. Term extraction, specially multiword terms is very important to help the creation of terminologies, ontologies and the improvement of machine translation. In this work we use a comparable corpora Portuguese/ English and want to find terms and their equivalents in both languages. To do this we start with separate term extraction for each language. Using morphossintatic patterns to identify n-grams (sequences of n words) most likely to be important terms of the domain. From the terms of each language, we use their context, i. e., the words that occurr around the term to compare the terms of different languages and to find the bilingual equivalents. We had as main goals in this work identificate monolingual terms, apply alignment techniques for Portuguese and evaluate the different parameters of size and type (used PoS) of window to the context extraction. This is the first work to apply this methodology to Portuguese and in spite of the lack of lexical and computational resources (like bilingual dictionaries and parsers) for this language, we achieved results comparable to state of the art in French/English.

This research shows and analyses how pupils in ninth grade in the nine-year compulsory school look at being multilingual, at the Swedish language and at Swedish as a second language. I also want to know if language and culture knowledge are taken care of in school as it says in official documents for the Swedish school and as it also is desired in an intercultural education. For my research I do eight interviews in a suburb school south of Stockholm. My work includes a limited account for the current research about being multilingual, Swedish as a second language and about the close relationship between language and culture. The applied theory is a social constructivist theory which in short can be explained like that we construct social structures that have consequences on the individual as well as on the society. I show in my research how the view of the world around the pupils influences their view on their languages, their school and their neighbourhood. I also draw conclusions from my interviews that language knowledge is little appreciated in school and that Swedish as a second language has a low status.

Este trabalho investiga técnicas de extração de termos multipalavra a partir de corpora comparáveis, que são conjuntos de textos em duas (ou mais) línguas sobre o mesmo domínio. A extração de termos, especialmente termos multipalavra é muito importante para auxiliar a criação de terminologias, ontologias e o aperfeiçoamento de tradutores automáticos. Neste trabalho utilizamos um corpus comparável português/inglês e queremos encontrar termos e seus equivalentes em ambas as línguas. Para isso começamos com a extração dos termos separadamente em cada língua, utilizando padrões morfossintáticos para identificar os n-gramas (sequências de n palavras) mais prováveis de serem termos importantes para o domínio. A partir dos termos de cada língua, utilizamos o contexto, isto é, as palavras que ocorrem no entorno dos termos para comparar os termos das diferentes línguas e encontrar os equivalentes bilíngues. Tínhamos como objetivos principais neste trabalho fazer a identificação monolíngue de termos, aplicar as técnicas de alinhamento para o português e avaliar os diferentes parâmetros de tamanho e tipo (PoS utilizados) de janela para a extração de contexto. Esse é o primeiro trabalho a aplicar essa metodologia para o Português e apesar da falta de alguns recursos léxicos e computacionais (como dicionários bilíngues e parsers) para essa língua, conseguimos alcançar resultados comparáveis com o estado da arte para trabalhos em Francês/Inglês. / This work investigates techniques for multiword term extraction from comparable corpora, which are sets of texts in two (or more) languages about the same topic. Term extraction, specially multiword terms is very important to help the creation of terminologies, ontologies and the improvement of machine translation. In this work we use a comparable corpora Portuguese/ English and want to find terms and their equivalents in both languages. To do this we start with separate term extraction for each language. Using morphossintatic patterns to identify n-grams (sequences of n words) most likely to be important terms of the domain. From the terms of each language, we use their context, i. e., the words that occurr around the term to compare the terms of different languages and to find the bilingual equivalents. We had as main goals in this work identificate monolingual terms, apply alignment techniques for Portuguese and evaluate the different parameters of size and type (used PoS) of window to the context extraction. This is the first work to apply this methodology to Portuguese and in spite of the lack of lexical and computational resources (like bilingual dictionaries and parsers) for this language, we achieved results comparable to state of the art in French/English.

Made available in DSpace on 2016-02-26T14:51:51Z (GMT). No. of bitstreams: 1 LuisBenedetti.pdf: 3716531 bytes, checksum: 705dd1ca707ede1f98a4da297f97497b (MD5) Previous issue date: 2012-05-30 / This research aims a discussion about the insertion of the digital information and communication technologies in the Act-do of educators and teachers dedicated to teaching of languages, primarily on the complex bi or multilingual contexts, work on the use of the web sites called blogs, or virtual diaries, as a pedagogical technique. This is the fi rst step to futures initiatives intending to develop a methodology, which s the fi rst product, was already delivered back to the community, the class: Teaching on multicultural and multilinguistic scenes directed to indigenous teachers from the city of Dourados, Mato Grosso do Sul, in particular the bilingual speakers working in the indigenous schools of the municipality. Initiate in a contextualization about the contemporary society, placed into the information age, or postmodernity, to the construction, helped by theoretical contributions from bilingualism, of a way for the use of the blogs in the school routine. Brings also the discussion held with the participating teachers about the importance of the blogs in the indigenous school, by its ability to provide transparency and to help the school to penetrate in the community, encouraging the school to become to be a space for the Guarani Nhandeva/Kaiowá and Terena s language/ culture rescue and promotion. Conclude pointing about the interest generated in the group of indigenous teachers participating, as well as the feedback given by then to the questions made, placing a list of the blogs built by these teachers during their participation in the course. / Este trabalho traz uma discussão sobre a inserção das tecnologias digitais da informação e da comunicação no agir-fazer pedagógico de educadores e professores dedicados ao ensino-aprendizagem de línguas, especialmente em contextos bi ou multilíngues complexos. Está focado na utilização dos mini sítios de internet chamados de blogs, ou diários virtuais, como ferramenta pedagógica. Trata-se do ponto de partida para novos estudos de uma metodologia cujo primeiro produto já foi entregue à comunidade, o curso: A formação do professor de cenário plurilíngue e pluricultural , direcionado a professores indígenas de Dourados, Mato Grosso do Sul, preferencialmente falantes bilíngues, que trabalhem nas escolas das aldeias do município. Parte de uma contextualização da sociedade contemporânea dentro da era informacional, ou pós- -modernidade, para a construção, a partir de aportes teóricos do bilinguismo, de uma proposta de aplicação dos blogs no cotidiano escolar. Traz a discussão realizada com os professores participantes sobre a importância do blog na escola indígena, sua capacidade de dotá-la de transparência e de capacidade de penetração na comunidade, para que o espaço escolar também se torne um espaço de resgate da língua e da cultura Guarani Ñhandeva/Kaiowá e Terena. Conclui apontando o interesse gerado no grupo de professores indígenas participantes, bem como o retorno dado por estes professores aos questionamentos levantados, trazendo uma lista dos blogs construídos pelo grupo durante sua participação no curso.

Sverige är ett mångkulturellt samhälle med många flerspråkiga klassrum. Eleverna i dessa klassrum har olika förhållanden till och erfarenheter av det språk som undervisningen, inklusive matematikundervisningen, sker på. Syftet med denna systematiska litteraturstudie är att synliggöra och sammanställa tidigare forskningsresultat om huruvida språket påverkar andraspråkselevers lärande i matematik. Elever i grundskolan (6-16 år) är av intresse för denna studie. Frågan som undersöks är Vad säger forskning om språkets påverkan för andraspråkselevers lärande i matematik? I resultatet synliggörs även vilka språkliga faktorer som gynnar, respektive missgynnar andraspråkselevers lärande i matematik. Studiens resultat är baserat på internationella artiklar. Resultatet visar på att språket påverkar andraspråkselevers lärande i matematik. De språkliga faktorerna som påverkar lärandet ligger hos eleven eller i undervisningen. / Sweden is a multicultural society with many multilingual classrooms. The pupils in these classrooms have different relationships with and experiences of the language of instruction, including mathematics teaching. The aim of this systematic literature study is to visualize and compile previous research findings about whether the language affects second language learners, learning in mathematics. Students in elementary school (6-16 years) are of interested in this study. The question under investigation is What does research say about the influence of language on language learning in mathematics? The result also shows which linguistic factors are beneficial and disadvantageous to the second language learners learning in mathematics. The results of the study are based on international articles. The result shows that language affect second language learners learning in mathematic. The linguistic factors that affect pupils learning exists either within the pupil or within the teaching.