Reproduce “Random geometric graphs with smooth kernels: sharp detection threshold and a spectral conjecture” 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
Cheng Mao, Yihong Wu, Jiaming Xu
Researchers from Georgia Institute of Technology, Yale University, and Duke University established the first sharp detection thresholds for random geometric graphs with smooth kernels. Their work introduced a novel posterior analysis technique for lower bounds and proposed a spectral conjecture unifying detection limits, finding that the critical dimension for detection is n^(3/4) for fixed kernels, while latent point recovery is possible at d = √n.
Key insights
- Detecting RGGs with smooth kernels is generally more challenging than detecting 'hard' RGGs, indicated by a lower critical dimension of n^(3/4) compared to n³.
- The ability to recover latent points (at d = √n) occurs at a substantially lower dimension than the ability to simply detect the geometric structure (at d = n^(3/4)).
- A unifying spectral conjecture, n³ tr²(κ³) = 1, based on the kernel operator, consistently predicts detection thresholds across diverse RGG models and kernel types.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.