Reproduce “AI agents can redesign engineering parts with more than a thousand CAD parameters and nonlinear simulations” 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
Alexandra Starostina, Alexander Hermann, Mahish Guru, Kartik Bali, Roland Aydin, Christian C. Cyron
The authors use an LLM to redesign a magnesium stent through simulation feedback, editing about 1,650 parametrised values while a separate pipeline builds and simulates proposals. In a clean run, the best valid design reduced the strain objective by 32%, from 0.1604 to 0.1085, within 15 simulations. One pilot design’s higher-fidelity check does not establish reliability of the reduced model.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.