A network information retrieval method based on pre-classification and feature learning noise resistance

A network information and feature learning technology, which is applied in the fields of information processing, network image search, machine learning, text comparison and network public security, can solve problems such as not considering data distribution information, and achieve the effect of real-time retrieval

Pending Publication Date: 2019-04-30
天罡网(北京)安全科技有限公司
View PDF4 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Although the content-based network information retrieval methods have achieved great results, these methods do not consider the distribution information of the data when extracting the characteristics of the network information. Information retrieval is very important

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
  • A network information retrieval method based on pre-classification and feature learning noise resistance
  • A network information retrieval method based on pre-classification and feature learning noise resistance
  • A network information retrieval method based on pre-classification and feature learning noise resistance

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0045]Below in conjunction with specific embodiment and accompanying drawing, the technical scheme of the present invention is described in more detail, and the flow chart of operation is as follows: figure 1 shown. The following examples are implemented on the premise of the technical solution of the present invention, and detailed implementation methods and processes are given, but the scope of protection of the present invention is not limited to image information retrieval in the embodiments, and can also be used for text information retrieval. When used for text retrieval, after the text data is preprocessed by cleaning and word segmentation, the words are embedding to form a low-dimensional dense word vector, which is used as the input of the deep neural network model.

[0046] A specific embodiment of the invention: The Pet Dataset image database. This is a 37-category pet dataset with 200 images for each category. Images vary widely in scale, pose, and lighting. The...

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 anti-noise network information retrieval method based on pre-classification and feature learning can be used for rapid retrieval of network information including text and image information. Firstly, a deep learning training network is used, a 16-layer deep neural network of VGG16 is selected, a rectification linear unit is used as an activation function, and a differential pressure layer is added after each complete connection layer. The number of nodes of the last completely connected layer is N, and a softmax function is used as a classification function of the last completely connectedlayer. After the network is trained, the data set extracts features in a forward propagation mode through the training network, a softmax is used for obtaining a pre-classification result from the last layer, and the data pre-classification result provides feedback information, namely the features belonging to the same category of network data should be stored together. And for the new query data,the similarity is measured between the query information and the information belonging to the same category in the query by using the cosine distance. The method has the advantages of high retrievalefficiency and high anti-noise performance.

Description

technical field [0001] The invention relates to a network information retrieval method based on pre-classification and feature learning and anti-noise, which can be widely used in the fields of network image search, text comparison, network public security and the like. It belongs to the field of machine learning and information processing. Background technique [0002] With the rapid development of Internet technology and cloud technology in recent years, network information has grown rapidly in an explosive manner. How to quickly and effectively retrieve the network information that users care about from these massive data has become a problem for researchers at home and abroad in recent years. popular research directions. Network information retrieval technology can be widely used in medical information retrieval, search engines, network security monitoring and other fields closely related to people's livelihood. [0003] At present, there are many methods for network i...

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
Patent Type & Authority Applications(China)
IPC IPC(8): G06F16/9535G06N3/04G06N3/08
CPCG06N3/08G06N3/045
Inventor 潘颋璇王斌
Owner 天罡网(北京)安全科技有限公司
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap