HGMP combines graph-level contrastive pre-training, type-aware graph augmentation, and per-node-type multiplicative prompts to improve few-shot node, edge, and graph classification on heterogeneous graphs.
A survey on knowl- edge graphs: Representation, acquisition, and applica- tions
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
-
HGMP:Heterogeneous Graph Multi-Task Prompt Learning
HGMP combines graph-level contrastive pre-training, type-aware graph augmentation, and per-node-type multiplicative prompts to improve few-shot node, edge, and graph classification on heterogeneous graphs.