DeepSDM 2025

DeepSDM (Deep Species Distribution Models) are a modern approach to Species Distribution Modeling (SDM) that leverages deep learning techniques to predict species distributions based on environmental and occurrence data. They extend traditional SDM methods, such as MaxEnt or GLMs, by using deep neural networks to model complex, non-linear relationships between species occurrences and environmental variables. As part of this symposium, we will be welcoming some of the leading specialists in these emerging methods to discuss the latest advances, as well as their potential applications and limitations. 

You will find on this page the replay of the sessions from DeepSDM 2025, on May 13.

Sponsors

Videos

Speakers

---

Organizers

Agenda

9:00 – 09:15 – Opening – Alexis Joly

9:15 – 10:45 – Morning Session 1  – Chair: Alexis Joly (Inria)

9:15-10:00 Devis Tuia (EPFL)
10:00-10:45 Florian Hartig (University of Regensburg)

11:15 – 12:30 Morning Session 2 – Chair:  Pierre Bonnet (Cirad, Pl@ntNet)
11:15 – 12:00 Phillip Brun & Benjamin Deneu (WSL) 

12:00 – 12:30 Damaris Zurell  (Potsdam University) – Visioconference

*************** LUNCH ***************

14:00 – 15:30 Afternoon Session 1 – Chair: David Mouillot

14:00 – 14:45 Sarah Si-Moussi & Selene Cerna ( LECA)
14:45 – 15:30 Alexis Joly & Diego Marcos (Inria, Pl@ntNet, University of Montpellier)

16:00 – 17:30 Round Table « What can deep learning offer to species distribution models for biodiversity knowledge and its conservation? » – (20 minutes of case study presentation followed by 1h10 of discussion) 

Chair: Christophe Botella (Inria), Diego Marcos (Inria)

Speakers: Clovis Grinand (Nitidae), David Mouillot (University of Montpellier), Alexis Joly (Inria), Philipp Brun (WSL), Laura Manocci (IRD), Olivier Gimenez (CNRS)

Location: 1000 Av. Agropolis, 34394 Montpellier