Eureka AIR delivers breakthrough ideas for toughest innovation challenges, trusted by R&D personnel around the world.

Adversarial network data generation method and abnormal event detection method

A technology of network data and abnormal events, applied in the application field of abnormal event detection, can solve the problems of low frequency and poor data generation effect, achieve the effect of reducing classification error and improving detection performance

Pending Publication Date: 2020-06-26
TSINGHUA UNIV
View PDF0 Cites 3 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

In particular, when there are small probability types of data in the data set, the relatively infrequent data generation effect is not good
In this regard, it is a possible breakthrough point to consider formulating an objective function based on a new type of information measure to optimize the data generation effect and improve the anomaly detection results of the generative adversarial network, and the relevant research is relatively rare at present.

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
  • Adversarial network data generation method and abnormal event detection method
  • Adversarial network data generation method and abnormal event detection method
  • Adversarial network data generation method and abnormal event detection method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0059] The method for generating adversarial network data in this embodiment uses exponential information measures to design the objective function of generating an adversarial network and its corresponding optimization method, thereby guiding the adversarial network to generate data.

[0060] Wherein, the equation of the objective function is:

[0061]

[0062] The corresponding optimization method is:

[0063]

[0064] Among them, D is the discriminator, G is the generator, and both the discriminator and the generator are neural networks; x and z are the inputs of the discriminator and the generator respectively; P r , P z Respectively, the probability distribution of the real data, the probability distribution of the data generated by the generator, and the probability distribution of the input data of the generator; Respectively represent the items in [] about the distribution P r , P z average calculation

[0065] Step 1: Experimental data preparation and ...

Embodiment 2

[0078] This embodiment is based on the abnormal event detection method against the network, including:

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 an adversarial network data generation method and an abnormal event detection method. The method is mainly designed for a multi-dimensional data set abnormal event detection scene in the field of current artificial intelligence and data mining. The generative adversarial network objective function and the corresponding optimization mode are desgined by using the exponentialinformation measure, thereby guiding the generative adversarial network to generate data. Furthermore, the conventional event data and the abnormal event data in the multi-dimensional data set are classified based on the exponential information measure generative adversarial network, so that the classification error is reduced, and the abnormal event detection performance is improved.

Description

technical field [0001] The present invention relates to the scene of abnormal event detection in the fields of artificial intelligence and data mining, in particular to the training of generating an adversarial network based on exponential information measurement and its application in abnormal event detection of multi-dimensional feature data sets. Background technique [0002] With the development of artificial intelligence technology and the advent of the era of big data, people's demand for how to mine more valuable information hidden in large amounts of data is becoming stronger and increasing day by day. Today, the development trend of data information processing technology has transformed from the original analysis and processing of low-dimensional isomorphic data to the mining and processing of high-dimensional heterogeneous data. In this regard, as one of the mainstream methods of current data information processing, artificial neural networks have been widely used ...

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 厍睿樊平毅刘善赟宛烁朱哲祺辛港涛
Owner TSINGHUA UNIV
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
Eureka Blog
Learn More
PatSnap group products