Reproduce “CoilStellaration: A Dataset and Benchmark for Engineering-Aware Stellarator Coilset Generation” 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
Jan-Hendrik Ewers, Santiago A. Cadena, Nicolò Foppiani, Andrea Merlo
CoilStellaration predicts candidate stellarator coilsets from target plasma boundaries and engineering requirements in a single forward pass, using optimized examples. On 1,289 tight-tier variable-shape instances, refinement from its learned start reached 22% strict feasibility after 200 iterations, versus 18% for REGCOIL after 500. This supports a limited-budget warm-start benefit, not faster end-to-end design or validation of final reactor coils.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.