Spectral Reversal: Counteracting Singular Value Bias for Graph Prompting
Pre-training Graph Neural Networks (GNNs) via self-supervised learning has become a dominant paradigm, yet efficiently adapting frozen encoders remains a challenge. Graph prompting offers a parameter-efficient alternative to fine-tuning, but existing methods largely treat pre-trained models as opaque feature extractors...
Han-Xu Yang, Yu-Huan Zhao, Xiao-Dong He et al.
· 0 citations