A hybrid label-and-feature propagation framework with attention and pseudo-label contrastive learning claims state-of-the-art node classification on graphs with partially missing features.
Terran Mott, Aaron Fanganello, and Tom Williams
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Oldie but Goodie: Re-illuminating Label Propagation on Graphs with Partially Observed Features
A hybrid label-and-feature propagation framework with attention and pseudo-label contrastive learning claims state-of-the-art node classification on graphs with partially missing features.