FasterGCN: Accelerating and enhancing graph convolutional network for recommendation

doi:10.1016/j.knosys.2026.115533

FasterGCN: Accelerating and enhancing graph convolutional network for recommendation

2026-02-10

0

PRE

AI

DOI：10.1016/j.knosys.2026.115533

Original

Original request for help

Save

Abstract

En 中文

• Potential Interaction Information (PII) is the key to accelerating and enhancing GCN. • A simple yet powerful quantitative function is designed to distill PII. • FasterGCN can effectively address under-smoothing and over-smoothing via PII. • FasterGCN achieves the best results in both efficiency and performance. • FasterGCN can be considered as a base model due to its relatively fixed parameters.

Keywords:

FasterGCN

Graph Convolutional Network

Potential Interaction Information

Under-smoothing

Over-smoothing

FasterGCN: Accelerating and enhancing graph convolutional network for recommendation

Abstract

Journal

Organization