MPCL: Multimodal Prompt Learning for Continual Relation Extraction with Type-Aware Inter-Modality Alignment

Xinyu Liu

doi:10.1016/j.inffus.2025.104025

MPCL: Multimodal Prompt Learning for Continual Relation Extraction with Type-Aware Inter-Modality Alignment

2025-12-03

0

PRE

AI

DOI：10.1016/j.inffus.2025.104025

Original

Original request for help

Save

Abstract

En 中文

• First propose prompt-based multimodal continual learning for MCRE without rehearsal • Attention-based joint matching of informative prompt knowledge for both modalities • Type-aware contrastive loss facilitates task-specific cross-modal semantic alignment

MPCL: Multimodal Prompt Learning for Continual Relation Extraction with Type-Aware Inter-Modality Alignment

Abstract

Journal

Organization