Information theory for hypergraph similarity | Science Advances
science.org
A new information-theoretic framework for comparing hypergraphs could reshape how we detect anomalies and cluster complex systems — where pairwise network models have always fallen short.
Information TheoryGraph TheoryComplexity TheoryNetwork Science
Theory Briefing
- Hypergraphs capture group interactions beyond pairs, but no principled similarity measure existed — until now.
- The new framework borrows from information theory to quantify how alike two hypergraphs are, enabling clustering and anomaly detection.
- Higher-order interactions are common in biology, social systems, and citations, making this tool broadly applicable across science.