Reproduce “Generative World Models Enable Predictive Control of Laser Melt Pool Dynamics” 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
Yiyang Yan, Markus Bambach, Mohamadreza Afrasiabi
The authors develop a differentiable world model that predicts laser powder-bed fusion melt-pool dynamics from temperature and phase histories under laser settings. It generates futures to support optimization of power and scan speed. Across 88 held-out transient cases, depth CRPS is 2.2 μm lower than a deterministic ViT. CFD replays show both controllers regulate depth on simulated overhangs, not on a printer.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.