Postdoctoral researcher at LISN. I am currently working on the automatic extraction of relevant entities from Clinical Trial Reports (CTRs), such as PICO entities, chemical, disease and drug names, … This project is supported by the CHIST-ERA grant CHIST-ERA-22-ORD-02, by the Luxembourg National Research Fund, by Swiss National Science Foundation, by the Agence Nationale de la Recherche, and by Engineering and Physical Sciences Research Council.

Interests

I’m interested in the nature of the exchanges between computer science and language science. Like many people, I think that these two disciplines have a lot to contribute to each other. This was one of the lines of thought of my PhD : Can we imagine that by a virtuous circle phenomenon, automated processing can refine the linguistic description at the same time as it allows to adjust the computer tools ?

My phd subject led me to study Multiword Expressions (MWEs) processing in NLP. I’m particularly interested in Unfrozen Multiword Expressions (UMWEs), which can be described as extreme cases of MWEs variations. What’s not to love about a process that can transform a famous recipe like duck confit into an asshole conflict, or notorious advertising slogans like eat 5 fruits and vegetables a day into eat 5 richs and vegetables a day ?

During my PhD, I worked on le Défricheur, a gamified annotation platform to help annotate UMWEs in a tweets dataset. As for now, this platform is only available in French. I also developed a tool for automatic unfrozen multiword identification in text: ASMR (Align, Segment, Match, Rank).