Artificial intelligence (AI) can enhance our understanding of biodiversity and the coexistence of plant species. This is demonstrated by a study led by the Desertification Research Centre (CIDE, UV-CSIC-GVA) in which transfer learning was employed to gain insights into plant species coexistence in regions of Murcia and Mexico, using data from a well-sampled ecosystem in Petrer, Alicante. The work, published in EcologicalInformatics, offers fresh perspectives on ecological interactions within plant communities and provides crucial information when limited data are available to address urgent ecological questions.
Transfer learning allows researchers to apply knowledge derived from large datasets to ecological communities for which less information is available.
"To understand how different plant species coexist in ecological communities, we trained an AI model using data from a well-studied community in southeast Spain, and then used it to predict interactions between species in two other communities-one in Spain and another in Mexico", explains Johannes Hirn, a CSIC researcher at CIDE and lead author of the study.
In addition to the CIDE team, researchers from the Institute of Corpuscular Physics (IFIC, UV-CSIC), the National Institute for Agricultural and Food Research and Technology (INIA-CSIC) and the National Autonomous University of Mexico contributed to the study. The team worked with plant communities in Petrer (Alicante), La Unión (Murcia) and San Juan Raya (Mexico), all’of them plant communities structured by facilitation interactions, that is, interactions between species that benefit at least one of the participants without causing harm to any of them.
The power of transfer learning
"In ecology, gathering field data is a slow and costly process, leading to many studies with small datasets", says Miguel Verdù, a CSIC researcher at CIDE and co-author of the study. "Here, we have shown that smaller datasets-with less than 1,000 vegetation patches, like those analysed in La Unión and San Juan Raya-can benefit from AI when combined with the larger dataset of more than 2,000 patches from the Petrer community, using transfer learning appropriately"."These techniques are just beginning to be used in basic ecology studies, but their development could help to improve restoration programmes for degraded areas or regions at risk of desertification", add José A. Navarro and Marta Goberna, from INIA, who also sign the paper.
According to the study, this advancement has significant implications for biodiversity conservation. It allows ecology to make better predictions about species coexistence and interactions using small datasets, thereby guiding ecological interventions more effectively.
The research highlights the role of artificial intelligence and deep learning neural networks in modelling complex species interactions more flexibly, providing a clearer picture of how they coexist across different environments.
"Our centre played a key role in developing the generative AI models used in this study as a basis for training and transferring data to different locations", says Verónica Sanz, professor of Physics at the University of Valencia, researcher at IFIC and co-author of the article in Ecological Informatics. "Much of our work focused on making the algorithm resilient to changes in species typical of each ecological environment, while remaining robust in the face of complex interactions", says the scientist.
The results suggest that transfer learning could become a standard tool in ecology, allowing researchers to use small datasets to address pressing ecological questions. Future studies could apply this technique to a broader range of ecosystems and species. "By transferring knowledge across ecosystems, we can begin to build a unified understanding of how patterns of species coexistence work", says Johannes Hirn. "This could allow us to make more informed conservation decisions", he concludes.
References :
J. Hirn, V. Sanz, J.E. García, et al., Transfer learning of species co-occurrence patterns between plant communities, Ecological Informatics (2024). DOI: https://doi.org/10.1016/j.ecoinf.2024.102826
CAPTION: Patterns of coexistence in well-sampled communities help to understand, thanks to AI, the coexistence of plants in ecologically close or distant communities. A. Montesinos, J. A. Navarro and A. Valiente-Banuet