Mathématiques et Informatique Appliquées
du Génome à l'Environnement

 

 

 

RALLM

Titre du projet
Robustness Assessment of LLM (Large Language Models) by designing evaluation metrics and by LLM experiments
Nom de l'appel d'offre
IB25
Agence de moyen
INRAE
Etat
Accepté
Année de soumission
2024
Equipe(s)
Bibliome
Coordinateur.trice
C. Nédellec
Participants de MaIAGE
C. Nédellec, R. Bossy, M. Courtin, X. Yao, A.-S. Foussat, X. Zhu
Partenaires (hors MaIAGE)
HZAU
Année de démarrage - Année de fin de projet
2025
Date de fin du projet
Résumé
The Bibliome research group and Jingbo Xia’s team share a common interest in ontology-based NLP methods aimed at extracting structured information from life science and agricultural documents to make textual data interoperable with other datasets. A critical challenge is assessing the extracted information based on its intended use, which goes beyond simply evaluating the performance of the methods. This project will focus on designing assessment rules that consider the quality of input data, the external knowledge applied, and the relevance of the extracted information in relation to external references. The work will be integrated into Bibliome’s ongoing NLP research for plant health monitoring, utilizing the EPOP annotated corpus and the EPPO knowledge base, with support from the biological experts involved in the ANR BEYOND project. Prof. Xia’s visit will also enhance collaboration with other NLP groups at Paris-Saclay University through seminars and meetings.