InfraredGP shows that a negative degree correction in a random-input spectral GNN produces clusterable embeddings, yielding fast, competitive graph partitioning without training.
Towards a better tradeoff between quality and efficiency of community detection: an inductive embedding method across graphs,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
InfraredGP: Efficient Graph Partitioning via Spectral Graph Neural Networks with Negative Corrections
InfraredGP shows that a negative degree correction in a random-input spectral GNN produces clusterable embeddings, yielding fast, competitive graph partitioning without training.