Research

Research conducted at CENTAL spans several complementary fields, at the intersection of natural language processing (NLP), artificial intelligence, and corpus linguistics.

NLP applied to language learning

Language learning is a particularly crucial issue in our multilingual and multicultural society. In order to promote this learning process (at a time when learners are under ever-increasing pressure), Computer Assisted Language Learning (CALL) opens up a number of perspectives by taking advantage of related disciplines: NLP, language pedagogy, readability, psycholinguistics, etc. Research in this field aims at developing new algorithms and new linguistic resources useful for the design of language learning or evaluation tools.

• Automatic assessment of written proficiency of French as a Foreign Language learners (FIDELIA)
• PhD thesis on automated corrective feedback in French as a Foreign Language
• CEFRLex: graded lexicons according to CEFR levels for several languages
• ReSyF: a graded synonym dictionary
• AMesure: a platform to support plain-language writing

Language models and deep learning

This research area focuses on deep learning methods and large language models themselves, considered both as objects of study and as tools used in several of the center's projects. It also includes methodological research specific to language models, particularly on the use of reinforcement learning to better represent the diversity of human judgments.

• PhD thesis on the diversity of human judgments in LLMs

Information extraction

CENTAL also specializes in developing information extraction applications and specialized semantic search engines for specific fields.

• BESOCIAL: towards a sustainable social media archiving strategy for Belgium
• Digital Literary Lab: archiving and preserving Belgian digital literature
• PhD thesis on automatic extraction of bilingual terminology from a French-English patent corpus

Analysis of digital social media

In the age of new media, this research area examines the linguistic and social skills we draw on to communicate, as well as the way language and communication practices evolve with new technologies, both in specialized sectors such as health and in intercultural contexts. The research combines qualitative and quantitative discourse analysis approaches, supported by NLP tools.

• PhD thesis on the analysis of online tourism discourse: a comparative, French-speaking and intercultural study of Belgian and Vietnamese tourism websites
• PhD thesis on the notion of lay expertise within online health communities