GeoDescriber x GeoLLM : Faire parler la donnée énergétique

HAL 2026 Open access

Read full text
Authors: Axel Declercq, Aude Lescouët, Jacqueline Tricot, Léon Lim, Eric Lavergne, Lydia Ould Ouali, Benoît Grossin
Cited by: 0
References: 5

Abstract

The GeoDescriber x GeoLLM project makes local energy data more understandable for communities. GeoDescriber automatically generates descriptive texts from territorial data, which GeoLLM then reformulates to enhance fluency and readability. This hybrid approach combines analytical rigor with editorial quality, making technical data intelligible for local stakeholders and showcasing the potential of language models to support the energy transition.

Fields of Study & Subjects

Topic Modeling Geographic Information Systems Studies Semantic Web and Ontologies

Publication details

Journal / source
HAL
Volume
Extraction et Gestion des Connaissances
Issue
RNTI-E-42
Pages
529–536
Publication date
27 January 2026
Language
FR