Document-level relation extraction method based on heterogeneous graph attention network
What is Al technical title?
Al technical title is built by PatSnap Al team. It summarizes the technical point description of the patent document.
A technology of relationship extraction and attention, applied in the field of extraction, can solve problems such as low accuracy
Pending Publication Date: 2022-06-21
HARBIN ENG UNIV
View PDF0 Cites 1 Cited by
Summary
Abstract
Description
Claims
Application Information
AI Technical Summary
This helps you quickly interpret patents by identifying the three key elements:
Problems solved by technology
Method used
Benefits of technology
Problems solved by technology
[0007] In order to solve the problem that the existing graph neural network ignores the nodes and edges in the graph when obtaining node representations, resulting in low accuracy of relation extraction, the present invention further proposes a document-level relation extraction based on a heterogeneous graph attention network method
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more
Image
Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
Click on the blue label to locate the original text in one second.
Reading with bidirectional positioning of images and text.
Smart Image
Examples
Experimental program
Comparison scheme
Effect test
specific Embodiment approach 1
[0056] Embodiment 1: Combining figure 1 Describe this embodiment, a method for document-level relationship extraction based on heterogeneous graph attention network described in this embodiment, which includes the following steps:
[0057] S1. Obtain document text;
[0058] First, given a document text, to predict the relationship between entities in the document text, and to facilitate the subsequent training of the document-level relationship extraction model.
[0059] S2, establish a document-level relationship extraction model, input the document text obtained in S1 into the document-level relationship extraction model for training, output the relationship of the document text, and obtain a trained document-level relationship extraction model. The specific process is:
[0060] The document-level relationship extraction model sequentially includes a vector representation layer, a context representation layer, a graph representation layer, and a classification layer;
[00...
Embodiment 1
[0113]Since different types of elements in a document play different roles in expressing semantic relations, the input document is constructed as a document graph with different node types, that is, the constructed document graph contains sentence nodes, mention nodes, and entity nodes. Then, seven types of undirected edges are constructed by exploiting the natural associations between document elements. Additionally, considering the importance of nodes and edges, a heterogeneous graph attention network is proposed to learn rich node representations in document graphs.
[0114] Specifically, given a document text as in, represents the dth in the document text x a word, d a =1,2,...i. At the same time, a document-level relationship extraction model is established, and the document-level relationship extraction model sequentially includes a vector representation layer, a context representation layer, a graph representation layer, and a classification layer.
[0115] docum...
Embodiment 2
[0171]
[0172]
[0173]
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more
PUM
Login to view more
Abstract
The invention relates to a document-level relation extraction method based on a heterogeneous graph attention network, in particular to a document-level entity relation extraction method based on the heterogeneous graph attention network, and aims to solve the problem of low relation extraction accuracy caused by neglect of nodes and edges in a graph when node representation is acquired by an existing graph neural network. The method comprises the following steps: S1, acquiring a document text; s2, establishing a document-level relation extraction model, inputting the document text obtained in the step S1 into the document-level relation extraction model for training, and outputting the relation of the document text to obtain a trained document-level relation extraction model; and S3, inputting the document text of which the document-level relationship is to be extracted into the trained document-level relationship extraction model in the S2 to obtain the relationship of the corresponding document text. Belongs to the technical field of computers.
Description
technical field [0001] The invention relates to an extraction method, in particular to a document-level entity relationship extraction method based on a heterogeneous graph attention network, and belongs to the field of computer technology. Background technique [0002] The relation extraction task can extract the semantic relation existing between two named entities in natural language text. Relationship extraction technology breaks through the traditional limitation of obtaining semantic relationships through manual reading and understanding. Instead, it is replaced by automatic search and extraction of semantic relationships, which can help computers better process text data and understand unstructured text. semantic information. At present, most of the related researches on relation extraction focus on the sentence level, that is, they only focus on the relationship between two entities within a sentence, and relatively little attention is paid to the situation across s...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more
Application Information
Patent Timeline
Application Date:The date an application was filed.
Publication Date:The date a patent or application was officially published.
First Publication Date:The earliest publication date of a patent with the same application number.
Issue Date:Publication date of the patent grant document.
PCT Entry Date:The Entry date of PCT National Phase.
Estimated Expiry Date:The statutory expiry date of a patent right according to the Patent Law, and it is the longest term of protection that the patent right can achieve without the termination of the patent right due to other reasons(Term extension factor has been taken into account ).
Invalid Date:Actual expiry date is based on effective date or publication date of legal transaction data of invalid patent.
Login to view more 
Login to view more