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

Theme text rapid detection method based on user intention embedded atlas learning

A technology of user intent and detection method, applied in the field of text detection, can solve problems such as relying on manual means, affecting detection efficiency, and low knowledge reuse rate, and achieve the effects of shortening response time, improving efficiency, and easy user operation

Pending Publication Date: 2022-05-20
国家计算机网络与信息安全管理中心上海分中心
View PDF0 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The current method is inaccurate and incomplete for the identification of network topic information, and relies heavily on manual means. Due to individual differences and lack of experience and knowledge, the current work mainly has problems such as low knowledge reuse rate and inaccurate and incomplete identification. Therefore, a method based on The topic text fast detection method of user intent embedding graph learning becomes particularly important
[0003] The existing fast topic text detection method based on user intention embedding graph learning has low algorithm recognition performance, which affects detection efficiency, and at the same time, the recognition accuracy and recall rate are low, which is inconvenient for manual operation; therefore, we propose a method based on user intention A Fast Topic Text Detection Method Based on Embedded Graph Learning

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
  • Theme text rapid detection method based on user intention embedded atlas learning
  • Theme text rapid detection method based on user intention embedded atlas learning
  • Theme text rapid detection method based on user intention embedded atlas learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0040] Reference Figure 1-4 , a rapid detection method of subject text based on user intent embedding map learning, the present embodiment specifically discloses a user preference and knowledge graph embedding method:

[0041] Identify topic text to extract topic features: Segment the subject text and identify subject entities, while extracting the subject feature keywords and converting them into word vectors.

[0042] Specific, such as Figure 2-3 As shown, first of all, through the Chinese Lexical Analysis System ICTCLAS Chinese word segmentation system of the Chinese Academy of Sciences, the subject text is segmented, part-of-speech annotation and subject entity recognition is carried out, and stop words and meaningless words are removed, and a group of keywords containing n describe the characteristics of the topic are obtained 1 ，w 2 ，...，w i ], and convert each set of feature keywords into word vectors, and map each group of word vectors to the corresponding d dimension repr...

Embodiment 2

[0047] Reference Figure 1-2 5, a rapid detection method of subject text based on user intent embedding map learning, in addition to the same structure as the above embodiment, the present embodiment specifically discloses a CNN model training method:

[0048] Structured processing of the knowledge graph: The entities and relationships related to the topic content knowledge graph and the topic features are obtained, and the TransD model is constructed to receive the relevant data, and the specific analysis is carried out for the topic description statement, and the subject embedded feature entity vector and the context entity vector are identified at the same time, so as to realize the knowledge graph embedding.

[0049] Specific, such as Figure 5 As shown, the TransD model is characterized by the text word w 1:i Perform entity similarity calculations with knowledge graph triplet candidate entities, eliminate ambiguity, obtain entity knowledge, and construct theme text knowledge su...

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 theme text rapid detection method based on user intention embedded atlas learning, and belongs to the technical field of text detection, and the detection method comprises the following specific steps: (1) recognizing a theme text to extract theme features; (2) embedding a user intention to perform structured extraction; (3) performing structured processing on the knowledge graph; (4) constructing a deep learning model to detect a subject text; according to the method, the user operation is easier, the method is close to the intention of the user, the method is closer to the manual judgment result, the algorithm recognition performance and efficiency can be improved by fusing the active search knowledge graph and the deep learning, and the recognition accuracy and the recall rate are high.

Description

Technical field [0001] The present invention relates to the field of text detection technology, in particular to a subject text rapid detection method based on user intent embedding pattern learning. Background [0002] Rapid perception of web-themed text content is particularly important in managing bad information on the Internet. In the context of big data, the network provides a hotbed for the rapid dissemination of information, bad information includes but is not limited to malicious dissemination, fabrication, resulting in damage to the country's reputation or interests of the information, this information is often with the theme, with certain characteristics, therefore, how to use knowledge and data, the use of empirical knowledge contained in the knowledge graph, combined with deep neural network model to identify and mine harmful information on the network is of great significance. The current method is not accurate to identify the network subject information is incomple...

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): G06F40/289G06F40/216G06F16/28G06N3/04G06N3/08
CPCG06F40/289G06F40/216G06F16/288G06N3/08G06N3/045
Inventor 刘丽娟闵宗茹郭承禹陈阳
Owner 国家计算机网络与信息安全管理中心上海分中心