AI and the World's Hard Problems · Earth & climate science · arXiv · 2026-10-20
Physics-anchored residual quantum neural networks for cloud-cover parameterization
The study applies compact quantum neural networks to cloud-cover parameterization, learning a correction on top of a fitted classical cloud-cover estimate. Across ten model variants on ICON/DYAMOND-derived data, residual prediction reduced test mean squared error relative to direct QNN prediction by up to 54.3%. A larger, longer-trained classical reference nevertheless achieved lower error, and the authors do not claim computational quantum advantage.