A survey on neural relation extraction
- PDF / 484,510 Bytes
- 19 Pages / 612 x 792 pts (letter) Page_size
- 59 Downloads / 316 Views
Print-CrossMark
October 2020 Vol. 63 No. 10: 1971–1989 https://doi.org/10.1007/s11431-020-1673-6
Special Topic: Natural Language Processing Technology
. Review .
A survey on neural relation extraction LIU Kang1,2* 1 National
Laboratory of Pattern Recognition, Institute of Automation, Chinese Academy of Sciences, Beijing 100190, China; 2 University of Chinese Academy of Sciences, Beijing 100049, China Received March 9, 2020; accepted June 9, 2020; published online September 15, 2020
Relation extraction is a key task for knowledge graph construction and natural language processing, which aims to extract meaningful relational information between entities from plain texts. With the development of deep learning, many neural relation extraction models were proposed recently. This paper introduces a survey on the task of neural relation extraction, including task description, widely used evaluation datasets, metrics, typical methods, challenges and recent research progresses. We mainly focus on four recent research problems: (1) how to learn the semantic representations from the given sentences for the target relation, (2) how to train a neural relation extraction model based on insufficient labeled instances, (3) how to extract relations across sentences or in a document and (4) how to jointly extract relations and corresponding entities? Finally, we give out our conclusion and future research issues. knowledge graph, relation extraction, event extraction and information extraction Citation:
Liu K. A survey on neural relation extraction. Sci China Tech Sci, 2020, 63: 1971–1989, https://doi.org/10.1007/s11431-020-1673-6
1 Introduction In a past decade, along with the prosperous progress of deep learning, knowledge engineering, as another important branch and fundamental infrastructure of artificial intelligence, has played more and more significant roles in many applications, such as natural language processing, information retrieval and recommendation. Through knowledge engineering, researchers or developers try to distill valuable information from explosive data and make them understandable for machine. In this way, artificial intelligence related applications could not be restricted on a shallow or superficial level and could have deep understanding of the data. Knowledge graph (KG), a recent well known form of knowledge, also called as knowledge base (KB), could provide structured semantic information about the complicated real world and has attracted a widespread attentions. Actu-
ally, KG could be regarded as a kind of structured and linked data, which stores the relational facts in a graph structure. In current KGs, the nodes usually denote entities (including entities, concepts, classes, properties etc.) and the edges/links between any two nodes represent their corresponding semantical relations. The basic factual unit in a knowledge graph is a triplet which is composed of two entities (nodes) and their semantic relation (edge). As shown in Figure 1, Chicago and United States are two entities, named as head
Data Loading...
