English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Recent studies attempted to utilize multilayer perceptrons (MLPs) to solvesemi-supervised node classification on graphs, by training a student MLP byknowledge distillation (KD) from a teacher graph neural network (GNN). Whileprevious studies have focused mostly on training the student MLP by matchingthe output probability distributions between the teacher and student modelsduring KD, it has not been systematically studied how to inject the structuralinformation in an explicit and interpretable manner. Inspired by GNNs thatseparate feature transformation $T$ and propagation $\Pi$, we re-frame the KDprocess as enabling the student MLP to explicitly learn both $T$ and $\Pi$.Although this can be achieved by applying the inverse propagation $\Pi^{-1}$before distillation from the teacher GNN, it still comes with a highcomputational cost from large matrix multiplications during training. To solvethis problem, we propose Propagate & Distill (P&D), which propagates the outputof the teacher GNN before KD and can be interpreted as an approximate processof the inverse propagation $\Pi^{-1}$. Through comprehensive evaluations usingreal-world benchmark datasets, we demonstrate the effectiveness of P&D byshowing further performance boost of the student MLP.

Unveiling the Unseen Potential of Graph Learning through MLPs: Effective Graph Learners Using Propagation-Embracing MLPs