Deep learning of fossil pollen morphology reveals 25,000 years of ecological change in East African grasslands
Authors: Adaime, M.-E., Kong, S., Urban, M. A., Street-Perrott, F. A., Verschuren, D., Punyasena, S. W.
Category: Plant Biology
▶ AI Summary
The study applies convolutional neural networks to super‑resolution images of modern and fossil grass pollen, enabling quantification of taxonomic diversity and C3/C4 physiological ratios in ancient grassland assemblages. Using semi‑supervised training and gradient‑boosted classifiers, the authors reconstruct past grass diversity and C4 decline in a 25,000‑year lake‑sediment record, linking these changes to climate and fire dynamics.