Reproduce “Neural Modal Decomposition: Architectural Priors from Observables” 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
Juho Park, Kaushik Sengupta
Neural Modal Decomposition predicts shared resonant modes from cavity geometry and estimates how queried ports couple to them, assembling impedance responses through a pole-residue representation. This permits changing port count under an open-termination condition. Trained only on two-port data, one model predicted held-out three-, four-, and five-port RF cavity responses, with five-port mean S-parameter MAE of 0.0445 ± 0.0570 on unseen configurations, in simulated pixelated RF.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.