Reproduce “FoundaRF: A Physics-Informed Foundation Model Framework for Radio-Frequency Passive Components” 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
Houbo He, Chris Han, Hyunsu Chae, Lei Xia, Qiang Zhou, Lawrence Liu, Sensen Li, David Z. Pan and 1 more
FoundaRF predicts frequency-dependent S-parameters from RF passive layouts using geometry pretraining and metadata-conditioned decoding. It is evaluated across four semiconductor technologies and magnetically coupled component families. Distance-field pretraining beats masked reconstruction in an SOI low-label test. For GaN, transfer reaches comparable accuracy with 25 labeled samples versus 100 from scratch. The results do not establish generalization beyond tested structures.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.