NEWS & EVENTS

We are proud to share that our team members, César Leblanc and Pierre Bonnet, presented their research at the International Botanical Congress in Madrid, from July 21-27, 2024.

César Leblanc introduced “GeoPl@ntBERT: Modelling Plant Species Assemblages and Producing High-Resolution Maps of Habitat Types with Large Language Models.” This novel framework uses deep learning to model plant species assemblages and produce high-resolution habitat maps. Combining computer vision and natural language processing, GeoPl@ntBERT classifies images and text to create detailed ecological models. Trained on millions of records from citizen scientists and vegetation surveys, it covers over ten thousand European vascular plant species, demonstrating high accuracy and supporting advanced biodiversity research.

Pierre Bonnet presented “Integrating Data for Conservation: A Use Case for Monitoring Plant Species on a European Scale.” His talk highlighted the Pl@ntNet platform’s role in enhancing biodiversity understanding across Europe. Leveraging citizen science and AI, Pl@ntNet identifies over 45,000 plant species and supports conservation efforts with high-resolution Species Distribution Models (DeepSDMs). These models help pinpoint key conservation areas and provide valuable data for management, monitoring, and promoting natural heritage at various scales.