As a crucial task in the field of natural language processing,document-level relation extraction aims to accurately ex-tract semantic relationships between entities from lengthy documents.Traditional document-level relation extraction methods ty-pically take the entire document as input.However,in reality,humans can predict relationships between entity pairs based on only a portion of the document,referred to as evidence sentences.In existing research,many methods start to utilize evidence sen-tences,but they face challenges such as incomplete evidence retrieval and difficulty in fully leveraging the advantages of these evi-dence sentences.To address this issue,we introduce a more efficient and accurate evidence sentence selection method.This is achieved by integrating a strategy for extracting evidence sentences through a fusion of formula-based and sentence-deletion-based approaches.We seamlessly integrate the evidence extraction with the training and inference processes,directing the document-le-vel relation extraction model to focus more on crucial sentences while still recognizing comprehensive information within the doc-ument.Experimental results demonstrate that the improved model outperforms existing models on public datasets.
NETL
NSTL
万方数据
