Abstract
Joint extraction of entities and their relations from the text is an essential issue in automatic knowledge graph construction, which is also known as the joint extraction of relational triplets. The relational triplets in sentence are complicated, multiple and different relational triplets may have overlaps, which is commonly seen in reality. However, multiple pairs of triplets cannot be efficiently extracted in most of the previous works. To mitigate this problem, we propose a deep neural network model based on the sequence-to-sequence learning, namely, the hybrid dual pointer networks (HDP), which extracts multiple pairs of triplets from the given sentence by generating the hybrid dual pointer sequence. In experiments, we tested our model using the New York Times (NYT) public dataset. The experimental results demonstrated that our model outperformed the state-of-the-art work, and achieved a 17.1% improvement on the F1 values.
Highlights
Joint extraction of entity mentions and their relations from unstructured text is an important task in information extraction (IE) and natural language processing (NLP)
Based on the joint decoding, a multiple relational triplets extraction model, i.e. the hybrid dual pointer networks is proposed in this paper
We are the first to employ the pointer for directly extracting entities from the inputs in the relational triplets extraction task
Summary
Joint extraction of entity mentions and their relations from unstructured text is an important task in information extraction (IE) and natural language processing (NLP). Wang et al [9] modelled entities and relations into the directed graph, and extracted multiple pairs of triplets by joint decoding. They assumed that overlapping edges cannot be existed in two identical nodes, so the directed graph modeling method cannot deal with the relation overlapping between two entities. Named HDP (Hybrid Dual Pointer model), see Fig. 2, which can jointly extract multiple pairs of triplets under the overlapping issue, and break the constraint that the multi-word entity cannot be extracted. We propose a hybrid dual pointer network model for joint extraction of the relational triplet. Experimental results on the NYT public dataset show that our model achieved the state-of-the-art performance in the multiple triplets extraction task
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
