.png){kind=icon}
.png){kind=icon}
Return
CITR: Context-driven implicit triple reasoning for joint multimodal entity-relation extraction
.png){kind=icon}
PRE
AI
DOI：10.1016/j.ipm.2025.104388.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• The JMERE task is reformulated as a type triple-centric sequence labeling problem. • LMMs-generated context enhances encoding, with constraints to avoid semantic bias. • Iteratively refines modality information related to potential type triples. • A dual-sequence tagging scheme reduces computation and label sparsity.
Journal
I
IF:
6.9
Papers:
5.2K
Citations:
1.4W