1
Return

MetaPFS: Memory-efficient node classification on text-attributed graphs via meta-guided progressive feature selection

delete2025-12-10
delete0
PRE
AI
Y
Yuewei Zhou
L
Lina Ni
Z
Zhijie Qu
X
Xuqiang Li
J
Jinquan Zhang
Y
Yongquan Liang
DOI：10.1016/j.ipm.2025.104542delete
deleteOriginal
deleteOriginal request for help
deleteShare
deleteSave
Abstract

Abstract

En 中文
• Identify feature redundancy as a major bottleneck in TAG node classification with PLMs. • Propose feature-agnostic warm-up training to inject global semantic priors before selection. • Propose progressive feature sharpening for smooth and consistent transition from soft to hard selection. • Propose a virtual-task-regularized meta learning policy to handle dynamic feature perturbations effectively. • Present MetaPFS, achieving competitive performance with minimal memory cost.

Journal

I
Information Processing and Management
IF:
6.9
Papers:
5.2K
Citations:
1.4W

Organization

S
Shandong University of Science and Technology
Scholars:
5.4K
Papers: 1.9K
Citations: 1.5W
Cited Papers

Cited Papers

Citing Papers

Citing Papers