Global ETD Search

11	Large-Context Question Answering with Cross-Lingual Transfer Sagen, Markus January 2021 (has links) Models based around the transformer architecture have become one of the most prominent for solving a multitude of natural language processing (NLP)tasks since its introduction in 2017. However, much research related to the transformer model has focused primarily on achieving high performance and many problems remain unsolved. Two of the most prominent currently are the lack of high performing non-English pre-trained models, and the limited number of words most trained models can incorporate for their context. Solving these problems would make NLP models more suitable for real-world applications, improving information retrieval, reading comprehension, and more. All previous research has focused on incorporating long-context for English language models. This thesis investigates the cross-lingual transferability between languages when only training for long-context in English. Training long-context models in English only could make long-context in low-resource languages, such as Swedish, more accessible since it is hard to find such data in most languages and costly to train for each language. This could become an efficient method for creating long-context models in other languages without the need for such data in all languages or pre-training from scratch. We extend the models’ context using the training scheme of the Longformer architecture and fine-tune on a question-answering task in several languages. Our evaluation could not satisfactorily confirm nor deny if transferring long-term context is possible for low-resource languages. We believe that using datasets that require long-context reasoning, such as a multilingual TriviaQAdataset, could demonstrate our hypothesis’s validity. Long-Context Multilingual Model Longformer XLM-R Longformer Long-term Context Extending Context Extend Context Large-Context Long-Context Large Context Long Context Cross-Lingual Multi-Lingual Cross Lingual Multi Lingual QA Question-Answering Question Answering Transformer model Machine Learning Transfer Learning SQuAD Memory Transfer Learning Long-Context Long Context Efficient Monolingual Multilingual QA model Language Model Huggingface BERT RoBERTa XLM-R mBERT Multilingual BERT Efficient Transformers Reformer Linformer Performer Transformer-XL Wikitext-103 TriviaQA HotpotQA WikiHopQA VINNOVA Peltarion AI LM MLM Deep Learning Natural Language Processing NLP Attention Transformers Transfer Learning Datasets Computer and Information Sciences Data- och informationsvetenskap
12	L'enseignement du français dans le Sud de l'Algérie. Du jeu théâtral à la production écrite dans une classe de 2e année de lycée à partir de Caligula d'Albert Camus / Teaching French in the southern Algeria. Of theatrical play to the production written in 2nd year of high school from Caligula of Albert Camus Oubah, Narimane 10 March 2017 (has links) Notre thèse envisage la production écrite en classe de français comme une pratique scolaire novatrice lorsqu’elle est accomplie au sein d’un espace-atelier et réalisée par l’entremise du jeu, de la mise en scène et du montage d’un spectacle. Pour ce faire, nous avons mis en œuvre une pièce de théâtre, Caligula d’Albert Camus, comme moyen d’enseignement-apprentissage, en l’accompagnant d’une gamme d’outils –dispositif ou unité didactique, vidéo de la pièce, carnet de bord, etc., – favorisant en parallèle une recherche-action sur terrain. Notre travail questionne en outre la possibilité d’une telle expérience menée dans un lieu peu commun, le Sud de l’Algérie, dans l’oasis de Bou-Saâda, et auprès d’un groupe d’apprenants d’une classe de 2e année Langues Étrangères au lycée. Nos objectifs sont de cultiver la compétence écrite chez l’apprenant oasien via l’écriture théâtrale, de lui donner goût à l’apprentissage de la langue et de conférer une certaine dynamique à cet apprentissage en lui permettant de créer et de voir représenté ce qu’il écrit sur le plateau. Au-delà d’une simple acquisition de compétences rédactionnelles, nous avons pu constater des mutations dans les représentations scolaires et sociales de la langue française chez le groupe-apprenant, issu d’une société sudiste conservatrice, chez qui la langue-cible a un statut différent de celui dont elle bénéficie au Nord, ce qui renvoie, par là même, à l’histoire, à la géographie et au plurilinguisme du pays. / Our thesis considers the written production, in a French class, as an innovative school practice when it is accomplished in a workshop space, and realized through the play, staging, and editing of a show. In order to do this, we have implemented a play, Caligula by Albert Camus, as a way of teaching and learning, accompanied by a range of tools - a teaching device or unit, a video of the play, Logbook, etc., - promoting, at the same time, an action research on the ground. Our work also questions the possibility of such an experiment conducted in an unusual place, the South of Algeria, in the oasis of Bou-Saâda, and with a group of learners 2nd Year, Foreign Languages in high school. Our aims are to improve the written skills of the oasis learner through theatrical scripture, to give him a taste to learn the language; and to give a certain dynamic to this learning by allowing him to create and see represented what 'He writes on the set’. Beyond a simple acquisition of editorial skills, we have seen changes in the academic and social representations of the French language in the learner group, from a conservative Southern society, in which the target language has a different status has the respect to the North, evoking the history, geography and multi-linguals of the country. Production écrite Écriture théâtrale Jeu théâtral Apprenant oasien Bou-Saâda Caligula Albert Camus Recherche-action Compétence écrite Sud de l’Algérie Plurilinguisme Written production Theatrical scripture Teaching and learning french LE Theatrical play Oasis learner Bou-Saâda Caligula Albert Camus Action research Written skills Southern Algeria Multi-lingual

Search results

Large-Context Question Answering with Cross-Lingual Transfer