DIGITAL.CSIC. Repositorio Institucional del CSIC
Publicación
Artículo científico (JournalArticle)
Comment on "Pollination supply models from a local to global scale": convolutional neural networks can improve pollination supply models at a global scale
Tools to predict pollinator activity at regional scales generally rely on land cover maps, combined with human-inferred mechanistic rules and/or expert knowledge. Recently, showed that, using large pollinator datasets, different environmental variables, and machine learning models, those predictions can be enhanced but at the cost of losing model interpretability. Here, we complement this work by exploring the potential of using advanced machine learning techniques to directly infer wild-bee visitation rates across different biomes only from land cover maps and available pollinator data while maintaining a mechanistic interpretation. In particular, we assess the ability of convolutional neural networks (CNNs), which are deep learning models, to infer mechanistic rules able to predict pollinator habitat use. At a global scale, our CNNs achieved a rank correlation coefficient of 0.44 between predictions and observations of pollinator visitation rates, doubling that of the previous human-inferred mechanistic models presented in (0.17). Most interestingly, we show that the predictions depend on both landscape composition and configuration variables, with prediction rules being more complex than those of traditional mechanistic processes. We also demonstrate how CNNs can improve the predictions of our previous data-driven models that did not use land cover maps by creating a new model that combined the predictions of our CNN with those of our best regression model based on environmental variables, a Bayesian ridge regressor. This new ensemble model improved the overall rank correlation from 0.56 to 0.64., This research was funded by the Comunidad de Madrid through the Research Grants for Young Investigators call from Universidad Politécnica de Madrid (APOYO-JOVENES-21-9K9EVJ-36-3VPZPJ). Alfonso Allen-Perkins was also funded by the Spanish Ministry of Science and Innovation (MICINN) through the TASTE (PID2021-127607OB-I00) and ChaSisCOMA (PID2021-122711NB-C21) projects. Ainhoa Magrach acknowledges funding from a Ministry of Science and Innovation grant (PID2021-127900NB-I00), an Ikerbasque research professorship, and the Spanish Ministry of Science and Innovation and the European Social Fund through the Ramón y Cajal Program (RYC2021-032351-I365). Research is co-funded by the European Union (ERC Consolidator Grant, GorBEEa 101086771). This research is part of the project PID2021-127900NB-I00 funded by the Ministry of Science and Innovation and supported by María de Maeztu Unit of Excellence 2023–2027, ref. CEX2021-001201-M, funded by MCIN/AEI /10.13039/501100011033., With funding from the Spanish government through the ‘María de Maeztu Unit of Excelence’ accreditation (CEX2021-001201-M), Peer reviewed
Proyecto:
AEI / Plan Estatal de Investigación Científica y Técnica y de Innovación 2021-2023 CEX2021-001201-M
,
AEI / Plan Estatal de Investigación Científica y Técnica y de Innovación 2021-2023 PID2021-127607OB-I00
,
AEI / Plan Estatal de Investigación Científica y Técnica y de Innovación 2021-2023 PID2021-122711NB-C21
,
AEI / Plan Estatal de Investigación Científica y Técnica y de Innovación 2021-2023 PID2021-127900NB-I00
DOI: http://hdl.handle.net/10261/388260, https://api.elsevier.com/content/abstract/scopus_id/85209880804
HANDLE: http://hdl.handle.net/10261/388260, https://api.elsevier.com/content/abstract/scopus_id/85209880804