The AI:EcoNet project aims to explore one of the grand challenges of modern ecology: predicting how species interactions in complex biological networks change across time and space. Species form intricate interaction networks (such as food webs) that are essential for ecosystem stability, yet these networks are increasingly disrupted by human activities, climate change, and biodiversity loss. Traditional models fall short of capturing this complexity, making it difficult to anticipate how ecosystems will respond to future pressures.

This project will pioneer Graph Representation Learning (GRL) methods tailored to ecological networks. By developing scalable models for signed, temporal, and attributed networks, we will reveal how interactions are lost, gained, or rearranged as species respond to shifting climates.