landmark-embedding

This paper analyzes distance-preserving embeddings in inhomogeneous random graphs, providing tighter distortion bounds than classical worst-case results and introducing a GNN-augmented variant that learns universal features from small graphs.