NML-GCL learns a negative metric network that down-weights false negatives during contrastive training, improving node classification and clustering on six standard graph benchmarks.
Towards expansive and adaptive hard negative mining: Graph con- trastive learning via subspace preserving
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Negative Metric Learning for Graphs
NML-GCL learns a negative metric network that down-weights false negatives during contrastive training, improving node classification and clustering on six standard graph benchmarks.