1,721,018 research outputs found
Les Modèles de Langage au Carrefour du Texte et de la Parole pour les Applications de Santé
No full textThe medical field presents unique language processing challenges through its specialized terminology, strict data regulations, and critical information needs. With the democratization of language models for assisting healthcare and clinical workers in their day-to-day work, the need for their adaptation to the domains of application became necessary to facilitate their accessibility to a broader audience, languages, and domains while reducing the computational cost of their usage.On the other hand, traditional approaches to medical speech processing rely on cascade systems that convert speech to text, apply natural language processing (NLP), and sometimes regenerate speech. While practical, these systems often lose paralinguistic features critical to clinical communication and suffer from error propagation between processing stages. Recent advances in self-supervised speech representation quantization have created new possibilities for integrating speech representation into other systems without intermediate text conversion, potentially preserving more communicative nuance.In this thesis, I investigate among other things, how speech capabilities can be integrated into existing text-based pre-trained language models with healthcare-related capabilities, leveraging their embedded medical knowledge while enabling direct speech processing. The examination of alignment between speech and text representations at various abstraction levels reveals potential pathways for effective cross-modal knowledge transfer with limited training data, a crucial consideration given healthcare's data constraints.Le domaine médical présente des défis uniques en matière de traitement du langage à travers sa terminologie spécialisée, ses réglementations strictes sur les données et ses besoins critiques en information. Avec la démocratisation des modèles de langage pour assister les professionnels de santé dans leur quotidien, leur adaptation aux domaines d'application est devenue nécessaire pour faciliter leur accessibilité à un public plus large, à différentes langues et domaines, tout en réduisant le coût computationnel de leur utilisation.D'autre part, les approches traditionnelles du traitement de la parole médicale reposent sur des systèmes en cascade qui convertissent la parole en texte, appliquent un traitement du language naturel (TAL), et parfois régénèrent la parole. Bien que pratiques, ces systèmes perdent souvent des caractéristiques paralinguistiques essentielles à la communication clinique et souffrent de la propagation d'erreurs entre les étapes de traitement. Les récentes avancées dans la quantification des représentations vocales auto-supervisées ont créé de nouvelles possibilités d'intégration de la représentation vocale dans d'autres systèmes sans conversion intermédiaire en texte, préservant potentiellement plus de nuances communicatives.Dans cette thèse, j'examine entre autre comment les capacités vocales peuvent être intégrées aux modèles de langage pré-entraînés basés sur le texte et possédant des connaissances liées aux domaines de la santé, en exploitant leurs connaissances médicales acquises tout en permettant un traitement direct de la parole, sans étapes intermédiaires. l'analyse des capacités d'alignement entre les représentations vocales et textuelles à différents niveaux d'abstraction ont révélé des méthodes plus optimales pour un transfert efficace de connaissances intermodales et savorisant ainsi l'apprentissage contraint par une quantité de données d'entraînement limitées, une considération cruciale étant donné les contraintes de données dans le domaine de la santé
Les Modèles de Langage au Carrefour du Texte et de la Parole pour les Applications de Santé
No full textThe medical field presents unique language processing challenges through its specialized terminology, strict data regulations, and critical information needs. With the democratization of language models for assisting healthcare and clinical workers in their day-to-day work, the need for their adaptation to the domains of application became necessary to facilitate their accessibility to a broader audience, languages, and domains while reducing the computational cost of their usage.On the other hand, traditional approaches to medical speech processing rely on cascade systems that convert speech to text, apply natural language processing (NLP), and sometimes regenerate speech. While practical, these systems often lose paralinguistic features critical to clinical communication and suffer from error propagation between processing stages. Recent advances in self-supervised speech representation quantization have created new possibilities for integrating speech representation into other systems without intermediate text conversion, potentially preserving more communicative nuance.In this thesis, I investigate among other things, how speech capabilities can be integrated into existing text-based pre-trained language models with healthcare-related capabilities, leveraging their embedded medical knowledge while enabling direct speech processing. The examination of alignment between speech and text representations at various abstraction levels reveals potential pathways for effective cross-modal knowledge transfer with limited training data, a crucial consideration given healthcare's data constraints.Le domaine médical présente des défis uniques en matière de traitement du langage à travers sa terminologie spécialisée, ses réglementations strictes sur les données et ses besoins critiques en information. Avec la démocratisation des modèles de langage pour assister les professionnels de santé dans leur quotidien, leur adaptation aux domaines d'application est devenue nécessaire pour faciliter leur accessibilité à un public plus large, à différentes langues et domaines, tout en réduisant le coût computationnel de leur utilisation.D'autre part, les approches traditionnelles du traitement de la parole médicale reposent sur des systèmes en cascade qui convertissent la parole en texte, appliquent un traitement du language naturel (TAL), et parfois régénèrent la parole. Bien que pratiques, ces systèmes perdent souvent des caractéristiques paralinguistiques essentielles à la communication clinique et souffrent de la propagation d'erreurs entre les étapes de traitement. Les récentes avancées dans la quantification des représentations vocales auto-supervisées ont créé de nouvelles possibilités d'intégration de la représentation vocale dans d'autres systèmes sans conversion intermédiaire en texte, préservant potentiellement plus de nuances communicatives.Dans cette thèse, j'examine entre autre comment les capacités vocales peuvent être intégrées aux modèles de langage pré-entraînés basés sur le texte et possédant des connaissances liées aux domaines de la santé, en exploitant leurs connaissances médicales acquises tout en permettant un traitement direct de la parole, sans étapes intermédiaires. l'analyse des capacités d'alignement entre les représentations vocales et textuelles à différents niveaux d'abstraction ont révélé des méthodes plus optimales pour un transfert efficace de connaissances intermodales et savorisant ainsi l'apprentissage contraint par une quantité de données d'entraînement limitées, une considération cruciale étant donné les contraintes de données dans le domaine de la santé
ANTILLES: An Open French Linguistically Enriched Part-of-Speech Corpus
No full textInternational audiencePart-of-speech (POS) tagging is a classical natural language processing (NLP) task. Although many tools and corpora have been proposed, especially for the most widely spoken languages, these suffer from limitations concerning their user license, the size of their tagset, or even approaches no longer in the state-of-the-art. In this article, we propose ANTILLES, an extended version of an existing French corpus (UD French-GSD) comprising an original set of labels obtained with the aid of morphological characteristics (gender, number, tense, etc.). This extended version includes a set of 65 labels, against 16 in the initial version. We also implemented several POS tools for French from this corpus, incorporating the latest advances in the state-of-the-art in this area. The corpus as well as the POS labeling tools are fully open and freely available
Team LIA/LS2N at BioCreative VII LitCovid Track: Multi-label Document Classification for COVID-19 Literature using Keyword Based Enhancement and Few-Shot Learning
No full textInternational audienceMulti-label text classification consists in attributing, for each textual document, one or more labels. Due to its nature, the task is often considered to be more challenging than other types of classification problems since the number of labels to assign is unknown. In text documents, this difficulty is generally the result of a blurry border between lexical fields of the labels or an underrepresentation of some of them. In this paper, we seek to automatically associate categories to scientific articles related to the COVID-19. We propose to address this multi-label classification problem by integrating an original keyword enhancement method to the TARS transformer-based approach designed to perform few-shot learning. Experiments conducted during the BioCreative challenge on the multi-label classification task show that our approach outperforms the baseline (ML-Net), no matter the metric considered
ANTILLES: An Open French Linguistically Enriched Part-of-Speech Corpus
No full textInternational audiencePart-of-speech (POS) tagging is a classical natural language processing (NLP) task. Although many tools and corpora have been proposed, especially for the most widely spoken languages, these suffer from limitations concerning their user license, the size of their tagset, or even approaches no longer in the state-of-the-art. In this article, we propose ANTILLES, an extended version of an existing French corpus (UD French-GSD) comprising an original set of labels obtained with the aid of morphological characteristics (gender, number, tense, etc.). This extended version includes a set of 65 labels, against 16 in the initial version. We also implemented several POS tools for French from this corpus, incorporating the latest advances in the state-of-the-art in this area. The corpus as well as the POS labeling tools are fully open and freely available
ANTILLES: An Open French Linguistically Enriched Part-of-Speech Corpus
No full textInternational audiencePart-of-speech (POS) tagging is a classical natural language processing (NLP) task. Although many tools and corpora have been proposed, especially for the most widely spoken languages, these suffer from limitations concerning their user license, the size of their tagset, or even approaches no longer in the state-of-the-art. In this article, we propose ANTILLES, an extended version of an existing French corpus (UD French-GSD) comprising an original set of labels obtained with the aid of morphological characteristics (gender, number, tense, etc.). This extended version includes a set of 65 labels, against 16 in the initial version. We also implemented several POS tools for French from this corpus, incorporating the latest advances in the state-of-the-art in this area. The corpus as well as the POS labeling tools are fully open and freely available
Going Beyond Counting First Authors in Author Co-citation Analysis
Full text linkThe present study examines one of the fundamental aspects of author co-citation analysis (ACA) - the way co-citation
counts are defined. Co-citation counting provides the data on which all subsequent statistical analyses and mappings
are based, and we compare ACA results based on two different types of co-citation counting - the traditional type that
only counts the first one among a cited work's authors on the one hand and a non-traditional type that takes into
account the first 5 authors of a cited work on the other hand. Results indicate that the picture produced through this non-traditional author co-citation counting contains more coherent author groups and is therefore considerably clearer. However, this picture represents fewer specialties in the research field being studied than that produced through the traditional first-author co-citation counting when the same number of top-ranked authors is selected and analyzed. Reasons for these effects are discussed
Variations on the Author
Full text link“Variations on the Author” discusses two of Eduardo Coutinho’s recent films (Um Dia na Vida, from 2010, and Últimas Conversas, posthumously released in 2015) and their contribution to the general question of documentary authorship. The director’s filmography is characterized by a consistent yet self-effacing form of authorial self-inscription: Coutinho often features as an interviewer that rather than express opinions propels discourses; an interviewer that is good at listening. This mode of self-inscription characterizes him as an author who is not expressive but who is nonetheless markedly present on the screen. In Um Dia na Vida, however, Coutinho is completely absent form the image, while Últimas Conversas, on the contrary, includes a confessional prologue that moves the director from the margins to the center of his films. This article examines the ways in which these works stand out in the filmography of a director who offers new insights into the notion of cinematic authorship
- …
