Our new research, “Inductive link prediction facilitates the discovery of missing links andenables cross-community inference in ecological networks”, is now available on Nature Ecology & Evolution!

The study introduces an inductive link-prediction framework trained on hundreds of ecological networks to infer missing species interactions in poorly sampled systems. By pooling data across plant–pollinator, host–parasite, plant–herbivore, and plant–seed-disperser communities, the model outperforms traditional approaches. We also provide a pre-trained online tool and Python package, enabling rapid predictions to guide ecological research and field validation.

Check out Shai’s blog about this publication: post

Read the full paper here: Nature Ecology & Evolution