Research Interests

• AI-assisted writing: helping researchers express their ideas clearly, with language models as assistants rather than authors.
  text revision · writing assistance · evaluation of revision systems

• AI-assisted verification of scientific content: making sure that claims rest on evidence, now that writing has become effortless and checking has not.
  claim verification · evidence in tables and figures · reliability of LLM-generated text

• Literature search: helping researchers find and make sense of an ever-growing literature.
  keyphrase generation · scientific document retrieval · paper recommendation · reading lists

Research Grants and Projects

As Principal Investigator

• 2026-2030 – FABULEUX project - Combating fabrications and misuse of LLMs in scientific writing
  Collaborative research project (PRC) funded by the French National Research Agency (ANR)
  This project aims to make LLM-assisted scientific writing more reliable by detecting and explaining fabricated or misleading content, and to develop methods and tools for identifying unreliable or fraudulent scientific publications.
  Collaborators: ISIR (Sorbonne Université, CNRS), ALMAnaCH team (Inria Paris), Mosaik team (LORIA, Université de Lorraine)

• 2023-2027 – NaviTerm project - Terminological navigation through the scientific literature
  Funded by the French Defense Innovation Agency (AID) and the CNRS
  This project aims to accelerate how researchers familiarize themselves with a new field by automatically generating navigable representations of scientific knowledge.
  Collaborators: Béatrice Daille (Nantes Université), Richard Dufour (Nantes Université)

• 2020-2025 – DELICES project - Indexing scientific literature through semantic expansion
  Young Researcher project (JCJC) funded by the French National Research Agency (ANR)
  This project aims to improve the retrieval effectiveness of search engines in scientific digital libraries by enriching their indexing with automatically generated keyphrases.
  Collaborators: Béatrice Daille (Nantes Université), Evelyne Jacquey (ATILF), Jian-Yun Nie (Université de Montréal)

• 2020 – WASP project – Building a writing assistance system for scientific papers
  Atlanstic2020 grant for a visiting researcher from the University of Tokyo (canceled by COVID-19)
  This project aims to design a writing assistance system that helps researchers select the most suitable formulaic expressions for their scientific papers.
  Collaborator: Akiko Aizawa (National Institute of Informatics, Japan)

• 2019 – IKEBANA project - Improving keyphrase extraction by adopting neural architectures
  Atlanstic2020 sabbatical grant, research stay at the National Institute of Informatics (NII), Tokyo
  This project aims to improve the effectiveness of automatic keyphrase extraction by using recurrent neural network models.
  Collaborator: Akiko Aizawa (National Institute of Informatics, Japan)

• 2016 – TALIAS project - Natural Language Processing for indexing scientific articles
  PEPS INS2I project funded by the CNRS
  This project aims to evaluate the robustness of keyword extraction models under varying levels of document noise.

• 2015 – GOLEM project - An Optimization-Based Approach for Keyphrase Extraction
  PEPS INS2I/INSMI project funded by the CNRS
  This project aims to explore the use of Integer Linear Programming methods for automatic keyword extraction.
  Collaborator: Evgeny Gurevsky (Nantes Université)

As Project Member

• 2012-2015 – TERMITH project - Terminologie et Indexation de Textes en sciences Humaines
  Funded by the French National Research Agency (ANR)
  This project aims to develop an automatic indexing method for scientific texts in the human and social sciences by leveraging domain-specific terms.