.png){kind=icon}
.png){kind=icon}
Return
Incomplete graph learning via data and representation-level interaction
.png){kind=icon}
PRE
AI
DOI：10.1016/j.knosys.2025.114583.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• An interactive framework is proposed to handle incomplete graphs. • Data interaction uses decoupled and refined steps for joint attribute-structure completion. • Representation interaction aligns three views via multi-view contrastive learning. • The framework beats 11 baselines on 8 datasets with 5.85 % average accuracy gain.