PhD Position in AI and Biodiversity Modeling

This PhD position is part of the PEPR DynaBiod initiative, hosted at Inria in Montpellier, France. The research focuses on advancing the understanding and prediction of plant community dynamics through scientific machine learning and mechanistic modeling. The successful candidate will design and evaluate hybrid artificial intelligence approaches, specifically Process-Informed Neural Networks (PINNs), to integrate ecological theory with observational data. Applicants are expected to work collaboratively in an interdisciplinary environment, conduct robust numerical experiments, and publish their findings in international scientific venues.

• Conduct a literature review on dynamic ecological system modeling, scientific machine learning, and PINNs.
• Explore mathematical formulations of ecological dynamics and integrate them into deep learning architectures.
• Design and develop PINN-based approaches and hybrid methods combining mechanistic models with machine learning.
• Investigate ecological parameter inference and spatiotemporal dynamics reconstruction using partial observational data.
• Develop simulation and evaluation tools using synthetic and real-world datasets such as vegetation inventories and remote sensing data.
• Assess model robustness, generalizability, and ecological relevance.
• Publish research results in international conferences and journals and contribute to open-source software development.