Unlock instant, AI-driven research and patent intelligence for your innovation.

Parameterized paper network node representing learning method

A learning method and network node technology, applied in neural learning methods, biological neural network models, data processing applications, etc., can solve the problem that new papers cannot perform representation learning and so on

Active Publication Date: 2018-06-29
BEIHANG UNIV +1
View PDF4 Cites 6 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] In order to solve the problem that new papers cannot be learned, the present invention proposes a parameterized paper network node representation learning 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
  • Parameterized paper network node representing learning method
  • Parameterized paper network node representing learning method
  • Parameterized paper network node representing learning method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0181] In this embodiment, the Cora paper data set and the Pubmed knowledge network data set are used for learning and experimental work.

[0182] Cora is a paper data set containing a total of 2708 paper nodes, including 2708 nodes and 5429 edges. Each node corresponds to a paper rich text information vector with a length of 1433. The rich text information vector is represented by 0 / 1 whether the word is exist. At the same time, each node is associated with a class attribute, and the total number of class attribute values ​​is 7.

[0183] Pubmed is a knowledge network dataset containing a total of 19,717 paper nodes, including 19,717 nodes and 44,338 edges. Each node corresponds to a paper-rich text information vector with a length of 500. The rich text information vector is represented by 0 / 1 word does it exist. At the same time, each node is associated with a class attribute, and the total number of class attribute values ​​is 3.

[0184] In order to verify the effective...

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

No PUM Login to View More

Abstract

The invention discloses a parameterized paper network node representing learning method. The method comprises the steps of firstly, constructing an idle paper node queue, and then adopting a random walk mode to sample neighbor nodes of any paper node and neighbor nodes of neighbors; regarding the selected paper nodes as first elements of the paper node queue, and afterwards, obtaining other elements of the paper node queue according to the skip probability; traversing all the paper nodes, and forming a paper node queue set; then adopting a positive and negative sampling method to generate neural network training data of a multi-layer perceptron; finally, adopting a neural network paper probability model for processing to obtain nonlinear transformation from paper node semantic informationto paper node vector representation, and then obtaining vector representation of the paper nodes.

Description

technical field [0001] The present invention relates to a representation learning method of a paper network, more particularly, a parameterized paper network node representation learning method. Background technique [0002] Social network belongs to the Internet concept noun. Social network service platforms such as Blog, WIKI, Tag, SNS, RSS, etc. The Internet has led to a brand-new human social organization and living mode quietly entering the present invention, and has built a huge group beyond the space of the earth—the network group. Human society in the 21st century is gradually emerging with new forms and characteristics. Individuals in the era of network globalization are aggregating into new social groups. The paper network refers to the networking of the relationship between papers and papers, which is manifested as mutual citations and shared authors between papers on the Internet. [0003] At present, the representation learning of the paper network mostly ado...

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
no application Login to View More
IPC IPC(8): G06F17/30G06N3/04G06N3/08G06Q50/00
CPCG06F16/3347G06F16/36G06N3/08G06Q50/01G06N3/048G06N3/045
Inventor 蒲菊华陈虞君刘伟班崟峰杜佳鸿熊璋
Owner BEIHANG UNIV