Moment-guided edge sampling

Connecting local edge edits to global graph structure through spectral moments.

Overview

Moment-guided edge sampling quantifies how local edge additions and removals alter global graph structure through spectral moments of the random-walk transition matrix. Exact combinatorial and low-rank updates give local edits interpretable structural signatures. The paper also studies preservation of related graph properties, including the triangle-weighted clustering coefficient, and applications to node classification and graph contrastive learning.

Illustration

Moment changes overview.

BibTeX

1. @article{cai2026moment,
     title = {Moment-guided edge sampling},
     author = {Cai, Weibin and Zafarani, Reza},
     journal = {arXiv preprint arXiv:2609.30472},
     year = {2026},
   }