Statut : Post-Doc
LLF, CNRS – UMR 7110
Université de Paris
Case 7031 – 5, rue Thomas Mann,
75205 Paris cedex 13
E-mail : email@example.com
Site Web : https://loicgrobol.github.io
This page not up-to-date. For more recent informations, please see https://loicgrobol.github.io.
I am a researcher in natural language processing, currently in a post-doc shared between Lattice at École Normale Supérieure and LLF at Université de Paris.
I received a PhD in Linguistics from Université Sorbonne Nouvelle in 2020 under the direction of Frédéric Landragin. The topic of my thesis was coreference resolution for spoken French and you can access it from HAL.
My research interests focus on machine learning applications to natural language processing; and specifically to irregular and low-ressource contexts, including spontaneous language, diachronic and historical data and low-resource languages. I am especially interested in probing limits and bias of machine learning models in these complex settings.
As for specific tasks, my main interests are syntactic parsing (particularly dependency syntax) and coreference resolution.
My current contract is with the Profiterole ANR project for which I work with Sophie Prévost and Benoît Crabbé on developping syntactic parsers for Old French, both from scratch and using tranfer learning techniques. I am also the maintainer of the SRCMF-UD corpus, a corpus of Old French in the Universal Dependencies metacorpus project.
In the past, I have also taught programming to NLP students, also in the PluriTAL master, as well as several classes in corpus linguistics and digital humanities at Université Sorbonne Nouvelle.
I am also a titular teacher of mathematics and computer science for the French secondary education system, although I am currently in leave for research activity.
A list of all my publications is available on my CV HAL