Graph-based manifold learning and diffusion processes provide a powerful framework for extracting intrinsic geometric features from high-dimensional data. By constructing a graph where nodes ...