Improving Expressive Power of Spectral Graph Neural Networks with Eigenvalue Correction

Eigenvalue correction enhances the expressive power of spectral graph neural networks by mitigating repeated eigenvalues and improving fitting capacity.