A Relation Extraction Method Based on Multi-layer Index and Cascading Binary Framework

Wanting Ji, Keyan Wen, Linlin Ding, Baoyan Song

Abstract: Relation extraction is a long-term and challenging task in information extraction, which aims to identify the entities from text and detect the relationships between the entities. Recently, relation extraction methods have been applied to many fields and achieved remarkable results. However, most of the existing relation extraction methods only focus on the relation extraction between general entities, resulting in low extraction efficiency and ignoring the prediction of nested entity relation types. To address this issue, this paper proposes a relation extraction method based on a multi-layer index and cascade binary framework named MICB. The proposed method enriches the entity information by establishing a multi-layer index to identify the nested entity in the sentence. A relation prediction module is designed to identify possible types of relationships in sentences, thereby improving the efficiency of relation extraction. The proposed method is tested on NYT and WebNLG relation extraction datasets. Experimental results show that the performance of the proposed method is superior to the existing relation extraction methods.