Computer Science · cs.CL
Double Graph Based Reasoning for Document-level Relation Extraction
Shuang Zeng, Runxin Xu, Baobao Chang, Lei Li
See on the mapAbstract
Document-level relation extraction aims to extract relations among entities within a document. Different from sentence-level relation extraction, it requires reasoning over multiple sentences across a document. In this paper, we propose Graph Aggregation-and-Inference Network (GAIN) featuring double graphs. GAIN first constructs a heterogeneous mention-level graph (hMG) to model complex interaction among different mentions across the document. It also constructs an entity-level graph (EG), based on which we propose a novel path reasoning mechanism to infer relations between entities. Experiments on the public dataset, DocRED, show GAIN achieves a significant performance improvement (2.85 on F1) over the previous state-of-the-art. Our code is available at https://github.com/DreamInvoker/GAIN .
References
The reference list for this paper has not been collected yet.
Cited by 1
1 paper in the index cites this (Semantic Scholar counts 290 in total).
The cited-by list is still filling in while reference lists are collected.