As a PhD student, you will investigate the potential of deep learning models for describing the suitability of habitats to host animal and vegetal species. In particular, you will develop models based on remote sensing data to model habitats where species can be found and prosper. To do so, you will leverage large-scale database of species distributions such as iNaturalist, eBird or Pl@ntNet and work towards integrated solutions involving a large number of species. The methods to be developed will be informed by knowledge in ecology, for instance about species interactions and species location priors. Such models will then be used to establish migration and extinction scenarios due to climate change projections.
Technical challenges related to large-scale prediction, data non-stationarity and quality of crowdsourced labels will be at the core of the studies.
Main duties and responsibilities include
- Perform data acquisition of large amounts of satellite data.
- Create a database of species observation data from multiple crowdsourced sources.
- Develop deep learning-based methodologies to analyze the data and create species distribution maps at scale, including ecological prior knowledge.
- Build and test scenarios of habitat suitability related to climate change.
- Write publications.
- Attend international conferences.
- Participate to education of the ENAC faculty.