Knowledge graph representation method based on graph convolutional network
A knowledge graph, convolutional network technology, applied in biological neural network model, unstructured text data retrieval, neural architecture and other directions, can solve the problems of unstable representation vector, difficult graph structure information representation, etc., to avoid noise.
Pending Publication Date: 2021-12-28
WUHAN UNIV
View PDF0 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0008] The purpose of the present invention is to provide a knowledge graph representation method based on a graph convolutional network to solve the problem in the above background technology that it is difficult to explicitly represent rich graph structure information; the representation vector of the relationship is unstable during the training process The problem
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 moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0040] Next, the technical solutions in the embodiments of the present invention will be described in the following examples in the embodiments of the present invention, and it is clear that the described embodiments are merely the embodiments of the present invention, not all of the embodiments. Embodiments in the present invention, those of ordinary skill in the art are in the range of protection of the present invention without making creative labor.
[0041] See figure 1 The present invention provides a technical solution: a knowledge map representation method based on the graph spacing network, the knowledge map representation method comprising the following steps:
[0042] Step 1: Random initialization relationship represents matrix;
[0043] Step 2: Random initialization relationship aggregation weight;
[0044] Step 3: Aggregation is represented by the relationship to the entity;
[0045] Step 4: The structure information encoding into the entity representation and relatio...
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 discloses a knowledge graph representation method based on a graph convolutional network. The knowledge graph representation method comprises the following steps: 1, randomly initializing a relation representation matrix; 2, randomly initializing a relation aggregation weight; 3, aggregating the relationship representation into an entity representation; 4, coding the structure information of the graph into entity representation and relation representation by utilizing graph convolution; and 5, performing translation model optimization training. In step 1, the relation representation matrix is randomly initialized, specifically, each behavior of the matrix Er is set to be a relation representation vector in the knowledge graph, random initialization is carried out, and the relation representation vector serves as the relation representation matrix. According to the knowledge graph representation method based on the graph convolutional network, the graph convolutional network is introduced to encode the graph structure, and graph structure information in a certain range is incorporated into a model consideration range, so that the graph structure information in the knowledge graph can be well captured; and a large amount of noise introduced by randomly initializing the entity representation is avoided through a method of aggregating the relationship representation into the entity representation.
Description
Technical field [0001] The present invention relates to the field of knowledge map related art, in particular, a method of known a graphic volume network representation. Background technique [0002] Knowledge maps expressed human knowledge in structured form, usually using the structure structure to represent the relationship between entities and entities in the real world (RELATION). Among them, the entity is a node in the figure structure, which is the side in the figure structure. Due to this structured expression of knowledge maps, it greatly enhances the ability of computer systems to express, storage, and reasoning. Since the 21st century, it has been widely used in various fields. Deep learning is a method of using computer imitating human neural network on information, and has made significant breakthroughs in image processing, natural language processing after the 1910s. In recent years, the knowledge of knowledge in knowledge spectrum is applied to the knowledge retrie...
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
Login to View More IPC IPC(8): G06F16/36G06N3/04
CPCG06F16/367G06N3/045
Inventor 刘泽彬
Owner WUHAN UNIV