Reproduce “ECO: Energy-Constrained Operator Learning for Chaotic Dynamics with Boundedness Guarantees” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Andrea Goertzen, Sunbochen Tang, Navid Azizan
ECO constrains neural-operator forecasts of chaotic fluid dynamics so long rollouts remain bounded. It projects each proposed state into a learned energy region, guaranteeing boundedness and convergence to the learned set under stated conditions. In Kuramoto–Sivashinsky tests, KL divergence falls from 0.6268 to 0.0208 versus unconstrained DeepONet; this empirical distributional agreement does not establish exact trajectory accuracy or recovery of true statistics.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.