RECA: Relation Extraction Based on Cross-Attention Neural Network
Xiaofeng Huang,Zhiqiang Guo,Jialiang Zhang,Hui Cao,Jie Yang
DOI: https://doi.org/10.3390/electronics11142161
IF: 2.9
2022-07-11
Electronics
Abstract:Extracting entities and relations, as a crucial part of many tasks in natural language processing, transforms the unstructured text information into structured information and provides corresponding data support for knowledge graph (KG) and knowledge vault (KV) construction. Nevertheless, the mainstream relation-extraction methods, the pipeline method and the joint method, ignore the dependency between the subject entity and the object entity. This work introduces a pre-trained BERT model and a dilated gated convolutional neural network (DGCNN) as an encoder to distinguish the long-range semantics representation from the input sequence. In addition, we propose a cross-attention neural network as a decoder to learn the importance of each subject word for each word of the input sequence. Experiments were undertaken with two extensive datasets, the New York Times Corpus (NYT) and WebNLG Corpus, and showed that our model performs significantly better than the CasRel model, outperforming the baseline by 1.9% and 0.7% absolute gain in terms of F1-score.
engineering, electrical & electronic,computer science, information systems,physics, applied