This paper claims the semi-supervised exact-recovery threshold for CSBM is I(aτ,bτ,cτ)>1 and that PCA, ridge regression on a linear GCN, and a trained GCN achieve it; the GCN proof is invalidated by an inconsistent estimator of the optimal self-loop weight.
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
REJECT 1representative citing papers
citing papers explorer
-
Optimal Exact Recovery in Semi-Supervised Learning: A Study of Spectral Methods and Graph Convolutional Networks
This paper claims the semi-supervised exact-recovery threshold for CSBM is I(aτ,bτ,cτ)>1 and that PCA, ridge regression on a linear GCN, and a trained GCN achieve it; the GCN proof is invalidated by an inconsistent estimator of the optimal self-loop weight.