Currently in PostDoc in the Laboratoire d'Ecologie Alpine, Grenoble, I'm interrested in the methodological development and applications of statistics to macro and community ecology. I'm motivated by monitoring biodiversity evolution with large and heterogeneous data from citizen sciences, and predicting effects of environmental changes on biodiversity to inform conservation planning.
My current CNRS PostDoc runs for 2 years for the project EcoNet. It focuses on studying methods for embedding ecological interactions networks into vector spaces in order to compare them, carry out multi-variate analysis and in the end understand how the architecture of interaction networks vary across space, environment or time.
I did my PhD INRAE at the UMR AMAP, Montpellier, France, where I studied statistical methods for species distribution models (SDM) based on large presence-only datasets coming from citizen sciences programs. It was done in close collaboration with the Pl@ntNet project, which motivated this work.
More precisely, this work included to (i) evaluate the benefits of deep learning and convolutional neural networks approaches for presence-only SDM, (ii) caracterize biases arising due to the distribution sampling effort, species niches and background points in presence only SDM based on Poisson point processes (iii) develop an new unbiased approach based on a joint model of sampling effort and species distributions, and (iv) measure the sampling and taxonomic coverage of Pl@ntNet contributions, in order to compare it with national botanical conservatories. This PhD was founded by a the national INRA-INRIA scholarship of 2016.
I am also co-organising GeoLifeCLEF (editions 2018, 2019 and 2020), a part of the LifeCLEF evaluation campaign. It is a machine learning challenge aiming at predicting the most likely species from geolocation.