A new machine learning tool developed by University of Michigan researchers can extract climate clues from digitized plant specimens in a fraction of the time traditional methods require. The approach could help scientists better understand today’s ecosystems and even reconstruct environments that existed millions of years ago.
