NOPE: Efficient Graph Coarsening via Non-Selfishness Principle
A new method for graph coarsening, named NOPE, has been proposed on arXiv. Graph coarsening reduces graph size while preserving structure and semantics. Existing methods use pair-wise similarity matching where each node independently seeks its best partner, causing high computational and memory costs. NOPE adopts a non-selfishness principle that prioritizes collective neighborhood interference, achieving linear memory and near-linear time complexity. A faster variant, NOPE*, reduces interference evaluation from O(δ·d) to O(d) using a local isotropy assumption, alleviating computational bottlenecks.
Key facts
- NOPE achieves linear memory consumption and near-linear computational complexity.
- NOPE* reduces interference evaluation from O(δ·d) to O(d).
- Existing graph coarsening methods rely on selfishness matching paradigm.
- NOPE uses non-selfishness principle prioritizing collective neighborhood interference.
- NOPE* is based on local isotropy assumption.
- Graph coarsening is a dimensionality reduction technique for graphs.
- The method addresses computational and memory overhead of existing approaches.
- The paper is available on arXiv with ID 2605.13021.
Entities
Institutions
- arXiv