Open-world semi-supervised relation extraction

Open Webpage

Diange Zhou, Yilin Duan, Shengwen Li, Hong Yao

Published: 01 Jan 2025, Last Modified: 12 Jun 2025Neural Networks 2025EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•Formally defines the task as Open-World Semi-Supervised Relation Extraction.•Propose a cluster-aware novel relation discovery method to adaptively capture the number of relation types.•Propose a novel SIND model that can extract mixed relation types in real-life scenarios.