Semi-automatic model updating system and model updating method

A model update and semi-automatic technology, applied in the field of Internet information security, can solve the problems of low data collection efficiency and accuracy, model update coverage and accuracy, etc., to improve the ability to identify abnormal data and strengthen the ability to learn and express , high accuracy and high efficiency

Pending Publication Date: 2020-11-20
WUHAN JIYI NETWORK TECH CO LTD
View PDF4 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] In view of the defects in the above-mentioned prior art, the purpose of the present invention is to provide a semi-automatic model update system and a model update method to solve the problem of low efficiency and accuracy of abnormal data collection in the prior art, resulting in poor coverage and accuracy of model update. low 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 more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Semi-automatic model updating system and model updating method
  • Semi-automatic model updating system and model updating method
  • Semi-automatic model updating system and model updating method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0044] see figure 1 As shown, a semi-automatic model update system includes a data preprocessing module, a suspicious data extraction module, a CNN model training module and a CNN model update module; the suspicious data extraction module includes a strategy marking module, a Cluster model marking module, and a manual marking module Module and Feature Visualization Markup Module.

[0045] Wherein, the data preprocessing module is used to perform standardized preprocessing on the original data to obtain a standardized data set, so as to improve the calculation efficiency of the data. Specifically include: (1) Determine the Schema of the data, and extract the main information from the userAgent, such as: device name, version, browser, version, etc. For example, the trajectory data is processed into a standard array, etc.; (2) the original row storage is converted into a column storage, reducing a large number of unnecessary repeated analysis; (3) pre-aggregation processing: suc...

Embodiment 2

[0063] see Figure 4 As shown, a semi-automatic model update system, compared with Embodiment 1, the difference is that the shown model update system also includes a suspicious data pattern division module, which is used to divide the pattern of the suspicious data set and eliminate Suspicious samples whose recognition degree of the original CNN model is lower than a preset threshold value are obtained to obtain several patterns of suspicious samples, so as to improve the recognition degree of suspicious samples in the original CNN model and reduce the risk of false sealing of the CNN model.

[0064] Further, the CNN model training module uses the suspicious samples of the several modes as training samples, and adopts the same method as that in Embodiment 1 to train the original CNN model to obtain a model training report and a new version of the CNN model.

Embodiment 3

[0066] see Figure 5 As shown, a semi-automatic model updating method adopts the semi-automatic model updating system described in Embodiment 1 to update the model, comprising the following steps:

[0067] S1. Create an original CNN model and deploy the original CNN model online;

[0068] S2. Perform standardized preprocessing on the original data to obtain a standardized data set;

[0069] S3. Regularly (7 days) extract and mark the suspicious data in the standardized data set through strategy marking, Cluster model marking, manual analysis marking and feature visualization marking of the original CNN model to obtain a suspicious data set;

[0070] S4. Using the suspicious data set as a training sample, train the original CNN model to obtain a model training report and a new version of the CNN model;

[0071] S5. According to the model training report, the data modeling engineer checks the recognition degree of the new version of CNN model for the suspicious samples and the f...

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 provides a semi-automatic model updating system and a model updating method. The method comprises steps of firstly, extracting and marking regularly suspicious data in a standardized data set through strategy marking, extracting and marking suspicious data in the standardized data set regularly to obtain a suspicious data set; performing mode division on the suspicious data set, andremoving data with low recognition degree; and then selecting an end-to-end CNN model for training and testing, and exporting a model training report and a visual interface of a training process. An engineer checks the recognition degree of the suspicious sample and the error sealing rate of the positive sample according to the model training report, so that the model precision is continuously improved, the error sealing rate is reduced, and an optimal new-edition CNN model is obtained. Suspicious data are acquired based on multiple angles and methods, and the CNN model is trained and updatedperiodically, so that the coverage rate and the recognition precision of the CNN model for abnormal mode data are gradually improved.

Description

technical field [0001] The invention belongs to the technical field of Internet information security, and in particular relates to a semi-automatic model updating system and a model updating method. Background technique [0002] With the rapid development of Internet information technology, the business scenarios applied by Internet technology are gradually increasing. While Internet applications bring convenience to users, they also bring certain risks. In order to reduce the risk of business processing and improve the security and controllability of the Internet, in the actual business processing, the server usually needs to identify the risks of the current business processing based on the preset model. [0003] Since Internet applications have the characteristics of fast scene changes and short update cycles, when updating models, not only must the updated models have high coverage and accuracy, but also require high model update efficiency. Before the model is updated...

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): G06N3/04G06N3/08
CPCG06N3/08G06N3/045
Inventor 雷炳盛陈国庆谢强
Owner WUHAN JIYI NETWORK TECH CO LTD
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