Reproduce “RF-GPT: Teaching AI to See the Wireless World” 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
Hang Zou, Yu Tian, Bohao Wang, Lina Bariah, Samson Lasaulce, Chongwen Huang, Mérouane Debbah
RF-GPT turns radio waveforms into spectrogram tokens that condition a language model to answer questions across several RF tasks. The system uses synthetic, standards-compliant signals and generated instruction-answer pairs for fine-tuning. On synthetic WBMC-Hard, it reaches 47.8% accuracy for exact modulation classification across 57 classes, versus 0.1% for its unmodified base model, but this does not establish real-world generalization.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.