Most graph neural network defenses improve explanation sparsity and stability in this benchmark, while the consistency and fidelity metrics largely saturate and stop being informative.
Explanations can be manipulated and geometry is to blame
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Robustness questions the interpretability of graph neural networks: what to do?
Most graph neural network defenses improve explanation sparsity and stability in this benchmark, while the consistency and fidelity metrics largely saturate and stop being informative.