Open-World Species Identification by Emulating Experts (eTaxonomist)

A €311,217 project (Inria) developing weakly supervised computer vision methods that emulate expert trait-based reasoning for zero-shot, interpretable species identification, validated on insects, birds, and plants (Pl@ntNet deployment). In collaboration with Alexis Joly (Inria, Pl@ntNet co-founder), Zeynep Akata (TU Munich / Helmholtz Munich), Dino Ienco (INRAe), Roberto Interdonato (Cirad), David Carral (Inria) and Marie-Laure Mugnier (Univ. Montpellier) on knowledge representation, Pierre Bonnet (Cirad, Pl@ntNet), Carole Kerdelhué (INRAe, insects), and Devin de Zwaan (Mount Allison University, birds), plus a PhD student and research engineer to be hired.