DeGLIF identifies noisy graph nodes by approximating how much each node's removal would improve loss on a small clean set, relabels them with the model's most confident alternative class, and retrains, improving accuracy under symmetric and pairwise label noise.
Learning Graph Neural Networks with Noisy Labels
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
abstract
We study the robustness to symmetric label noise of GNNs training procedures. By combining the nonlinear neural message-passing models (e.g. Graph Isomorphism Networks, GraphSAGE, etc.) with loss correction methods, we present a noise-tolerant approach for the graph classification task. Our experiments show that test accuracy can be improved under the artificial symmetric noisy setting.
citation-role summary
background 1
citation-polarity summary
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
DeGLIF for Label Noise Robust Node Classification using GNNs
DeGLIF identifies noisy graph nodes by approximating how much each node's removal would improve loss on a small clean set, relabels them with the model's most confident alternative class, and retrains, improving accuracy under symmetric and pairwise label noise.