Séminaire AMAP - Atelier technique


05/03/2024 de 14h00 à 15h00PS 2 salle 201

Recent advances in deep learning for ecology suggest that deep neural networks will offer significant advancements in modelling species distributions thanks to the increasing number of observations and high quality data: satellite, bioclimatic, timeseries… These days, the PyTorch framework is becoming increasingly popular to build such deep learning pipelines thanks to its completeness and huge contributing community.
However, ecologists do not necessarily have a strong background in either PyTorch, Python or even deep learning in itself which can be barrier to putting efforts in developping such methods.
In the context of the MAMBO european project, we are currently developping a Python framework, MALPOLON, to facilitate the use of neural networks to learn Deep-SDM (Species Distibution Models), even with minimal knowledge of the language.
We believe this tool could be useful for a number of researchers and we are looking for feedback and identifying more user needs.

Télécharger l'affiche