Search Results for author:
Found 46 papers, 14 papers with code
Mesures linguistiques automatiques pour l’évaluation des systèmes de Reconnaissance Automatique de la Parole (Automated linguistic measures for automatic speech recognition systems’ evaluation)
L’évaluation de transcriptions issues de systèmes de Reconnaissance Automatique de la Parole (RAP) est un problème difficile et toujours ouvert, qui se résume généralement à ne considérer que le WER.
Automatic Speech Recognition
Automatic Speech Recognition (ASR)
+1
Correction automatique d’examens écrits par approche neuronale profonde et attention croisée bidirectionnelle (Deep Neural Networks and Bidirectional Cross-Attention for Automatic Answer Grading)
Nous proposons trois systèmes de classification reposant sur des caractéristiques extraites de plongements de mots contextuels issus d’un modèle BERT (CamemBERT).
Remplacement de mentions pour l’adaptation d’un corpus de reconnaissance d’entités nommées à un domaine cible (Mention replacement for adapting a named entity recognition dataset to a target domain)
La reconnaissance d’entités nommées est une tâche de traitement automatique du langage naturel bien étudiée et utile dans de nombreuses applications.
A Benchmark of French ASR Systems Based on Error Severity
Automatic Speech Recognition (ASR) transcription errors are commonly assessed using metrics that compare them with a reference transcription, such as Word Error Rate (WER), which measures spelling deviations from the reference, or semantic score-based metrics.
Automatic Speech Recognition
Automatic Speech Recognition (ASR)
+1
ParaRev: Building a dataset for Scientific Paragraph Revision annotated with revision instruction
Revision is a crucial step in scientific writing, where authors refine their work to improve clarity, structure, and academic quality.
ACL-rlg: A Dataset for Reading List Generation
Familiarizing oneself with a new scientific field and its existing literature can be daunting due to the large amount of available articles.
The Role of Natural Language Processing Tasks in Automatic Literary Character Network Construction
The automatic extraction of character networks from literary texts is generally carried out using natural language processing (NLP) cascading pipelines.
Whole-Graph Representation Learning For the Classification of Signed Networks
In this article, we tackle this issue by proposing two approaches to learning whole-graph representations of general signed graphs.
Renard: A Modular Pipeline for Extracting Character Networks from Narrative Texts
Renard (Relationships Extraction from NARrative Documents) is a Python library that allows users to define custom natural language processing (NLP) pipelines to extract character networks from narrative texts.
Zero-Shot End-To-End Spoken Question Answering In Medical Domain
In the rapidly evolving landscape of spoken question-answering (SQA), the integration of large language models (LLMs) has emerged as a transformative development.
CASIMIR: A Corpus of Scientific Articles enhanced with Multiple Author-Integrated Revisions
Writing a scientific article is a challenging task as it is a highly codified and specific genre, consequently proficiency in written communication is essential for effectively conveying research findings and ideas.
Probing the Information Encoded in Neural-based Acoustic Models of Automatic Speech Recognition Systems
Following many researches in neural networks interpretability, we propose in this article a protocol that aims to determine which and where information is located in an ASR acoustic model (AM).
Automatic Speech Recognition
Automatic Speech Recognition (ASR)
+3
How Important Is Tokenization in French Medical Masked Language Models?
Subword tokenization has become the prevailing standard in the field of natural language processing (NLP) over recent years, primarily due to the widespread utilization of pre-trained language models.
DrBenchmark: A Large Language Understanding Evaluation Benchmark for French Biomedical Domain
This limitation hampers the evaluation of the latest French biomedical models, as they are either assessed on a minimal number of tasks with non-standardized protocols or evaluated using general downstream tasks.
Language Model Adaptation to Specialized Domains through Selective Masking based on Genre and Topical Characteristics
Recent advances in pre-trained language modeling have facilitated significant progress across various natural language processing (NLP) tasks.
BioMistral: A Collection of Open-Source Pretrained Large Language Models for Medical Domains
This marks the first large-scale multilingual evaluation of LLMs in the medical domain.
Ranked #8 on
Few-Shot Learning
on MedConceptsQA
Learning to Rank Context for Named Entity Recognition Using a Synthetic Dataset
Using this dataset, we train a neural context retriever based on a BERT model that is able to find relevant context for NER.
A Zero-shot and Few-shot Study of Instruction-Finetuned Large Language Models Applied to Clinical and Biomedical Tasks
We evaluate four state-of-the-art instruction-tuned large language models (LLMs) -- ChatGPT, Flan-T5 UL2, Tk-Instruct, and Alpaca -- on a set of 13 real-world clinical and biomedical natural language processing (NLP) tasks in English, such as named-entity recognition (NER), question-answering (QA), relation extraction (RE), etc.
The Role of Global and Local Context in Named Entity Recognition
Pre-trained transformer-based models have recently shown great performance when applied to Named Entity Recognition (NER).
FrenchMedMCQA: A French Multiple-Choice Question Answering Dataset for Medical domain
This paper introduces FrenchMedMCQA, the first publicly available Multiple-Choice Question Answering (MCQA) dataset in French for medical domain.
DrBERT: A Robust Pre-trained Model in French for Biomedical and Clinical domains
In recent years, pre-trained language models (PLMs) achieve the best performance on a wide range of natural language processing (NLP) tasks.
Text revision in Scientific Writing Assistance: An Overview
Writing a scientific article is a challenging task as it is a highly codified genre.
Data Augmentation for Robust Character Detection in Fantasy Novels
Named Entity Recognition (NER) is a low-level task often used as a foundation for solving higher level NLP problems.
ANTILLES: An Open French Linguistically Enriched Part-of-Speech Corpus
Part-of-speech (POS) tagging is a classical natural language processing (NLP) task.
Ranked #1 on
Part-Of-Speech Tagging
on ANTILLES
Multi-label classification for biomedical literature: an overview of the BioCreative VII LitCovid Track for COVID-19 literature topic annotations
To close the gap, we organized the BioCreative LitCovid track to call for a community effort to tackle automated topic annotation for COVID-19 literature.
Label Refining: a semi-supervised method to extract voice characteristics without ground truth
To address this problem, we propose a new semi-supervised learning method entitled Label Refining that consists in extracting refined labels (e. g. vocal characteristics) from known initial labels (e. g. character played in a recording).
Apprentissage automatique de repr\'esentation de voix \`a l'aide d'une distillation de la connaissance pour le casting vocal (Learning voice representation using knowledge distillation for automatic voice casting )
Les exp{\'e}riences men{\'e}es sur des extraits de voix de jeux vid{\'e}o montrent une am{\'e}lioration significative de l{'}approche p-vecteur, avec distillation de la connaissance, par rapport {\`a} une repr{\'e}sentation x-vecteur, {\'e}tat-de-l{'}art en reconnaissance du locuteur.
La voix act\'ee : pratiques, enjeux, applications (Acted voice : practices, challenges, applications)
La voix act{\'e}e repr{\'e}sente un d{\'e}fi majeur pour les futures interfaces vocales avec un potentiel d{'}application extr{\^e}mement important pour la transformation num{\'e}rique des secteurs de la culture et de la communication, comme la production ou la post-production de voix pour les s{\'e}ries ou le cin{\'e}ma.
A Multimodal Educational Corpus of Oral Courses: Annotation, Analysis and Case Study
The use cases concern scientific fields from both speech and text processing, with language model adaptation, thematic segmentation and transcription to slide alignment.
WAC: A Corpus of Wikipedia Conversations for Online Abuse Detection
This large corpus of more than 380k annotated messages opens perspectives for online abuse detection and especially for context-based approaches.
Abusive Language Detection in Online Conversations by Combining Content-and Graph-based Features
In recent years, online social networks have allowed worldwide users to meet and discuss.
Conversational Networks for Automatic Online Moderation
We identify the most appropriate network extraction parameters and discuss the discriminative power of our features, relatively to their topological and temporal nature.
Graph-based Features for Automatic Online Abuse Detection
While online communities have become increasingly important over the years, the moderation of user-generated content is still performed mostly manually.
Automatic Text Summarization Approaches to Speed up Topic Model Learning Process
One of the main difficulty in using topic model on huge data collection is related to the material resources (CPU time and memory) required for model estimate.
Systèmes du LIA à DEFT'13
The 2013 D\'efi de Fouille de Textes (DEFT) campaign is interested in two types of language analysis tasks, the document classification and the information extraction in the specialized domain of cuisine recipes.
Parallel Long Short-Term Memory for Multi-stream Classification
Nevertheless, these RNNs process a single input stream in one (LSTM) or two (Bidirectional LSTM) directions.
Un Corpus de Flux TV Annot\'es pour la Pr\'ediction de Genres (A Genre Annotated Corpus of French Multi-channel TV Streams for Genre Prediction)
Cet article pr{\'e}sente une m{\'e}thode de pr{\'e}diction de genres d{'}{\'e}missions t{\'e}l{\'e}vis{\'e}es couvrant 2 jours de diffusion de 4 cha{\^\i}nes TV fran{\c{c}}aises structur{\'e}s en {\'e}missions annot{\'e}es en genres.
Un Sous-espace Th\'ematique Latent pour la Compr\'ehension du Langage Parl\'e (A Latent Topic-based Subspace for Spoken Language Understanding)
Les applications de compr{\'e}hension du langage parl{\'e} sont moins performantes si les documents transcrits automatiquement contiennent un taux d{'}erreur-mot {\'e}lev{\'e}.
Auto-encodeurs pour la compr\'ehension de documents parl\'es (Auto-encoders for Spoken Document Understanding)
Les repr{\'e}sentations de documents au moyen d{'}approches {\`a} base de r{\'e}seaux de neurones ont montr{\'e} des am{\'e}liorations significatives dans de nombreuses t{\^a}ches du traitement du langage naturel.
Apport de l'information temporelle des contextes pour la repr\'esentation vectorielle continue des mots
Ces approches sont manipul{\'e}es au travers d{'}un r{\'e}seau de neurones, l{'}architecture CBOW cherchant alors {\`a} pr{\'e}dire un mot sachant son contexte, alors que l{'}architecture Skip-Gram pr{\'e}dit un contexte sachant un mot.
Initialisation de R\'eseaux de Neurones \`a l'aide d'un Espace Th\'ematique
La m{\'e}thode propos{\'e}e consiste {\`a} configurer la topologie d{'}un ANN ainsi que d{'}initialiser les connexions de celui-ci {\`a} l{'}aide des espaces th{\'e}matiques appris pr{\'e}c{\'e}demment.
Characterizing and Predicting Bursty Events: The Buzz Case Study on Twitter
The prediction of bursty events on the Internet is a challenging task.
A LDA-Based Topic Classification Approach From Highly Imperfect Automatic Transcriptions
Although the current transcription systems could achieve high recognition performance, they still have a lot of difficulties to transcribe speech in very noisy environments.
