Learning to Emulate Chaos: Adversarial Optimal Transport Regularization

arXiv:2604. 21097v2 Announce Type: replace Abstract: Chaos arises in many complex dynamical systems, from weather to power grids, but is difficult to accurately model with data-driven methods such as machine learning emulators.