Reproduce “Active learning of deep surrogates for PDEs: Application to metasurface design” 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
Raphaël Pestourie, Youssef Mroueh, Thanh V. Nguyen, Payel Das, Steven G. Johnson
The authors present an active-learning method that trains a neural-network surrogate for unit-cell optical responses, selecting new cells for costly simulations where it estimates high error. For a ten-parameter multilayer cell it reaches fractional error 0.07, with 12 times fewer training points than random sampling. A 100-cell multiplexer result is numerical and relies on the locally periodic approximation.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.