The abstract of this arXiv preprint argues that developing resilient crops requires models that capture how plant growth emerges from the continuous interplay between genetic and environmental factors. It notes that high-throughput phenotyping platforms generate data that could support such models, though the abstract is cut off before describing the full methodology.
The title indicates a method called "physiology-informed sparse identification." This suggests the authors integrate physiological principles into a sparse identification framework, which is typically used to discover the simplest dynamic equations that explain observed data. By constraining the search with known plant biology, the approach may yield models that are both data-driven and biologically interpretable.
As a preprint, the work has not yet undergone peer review. The available abstract does not include details on validation, performance, or comparison with existing crop models, so the claims should be treated as preliminary until the full paper is examined.