GHGRL uses LLM-generated type labels and confidence scores to drive a parameter-adaptive GNN, achieving strong heterogeneous graph classification accuracy without human-provided type information.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Bootstrapping Heterogeneous Graph Representation Learning via Large Language Models: A Generalized Approach
GHGRL uses LLM-generated type labels and confidence scores to drive a parameter-adaptive GNN, achieving strong heterogeneous graph classification accuracy without human-provided type information.