A node-level joint-embedding predictive architecture that masks k-hop ego-subgraphs and predicts latent targets achieves the best average rank among six self-supervised methods on five node classification benchmarks.
Proceedings of the 28th ACM SIGKDD Conference on Knowledge Discovery and Data Mining , pages=
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
NodeJEPA: Structure-Conditioned Latent Prediction for Node-Level Graph Self-Supervised Learning
A node-level joint-embedding predictive architecture that masks k-hop ego-subgraphs and predicts latent targets achieves the best average rank among six self-supervised methods on five node classification benchmarks.