Safety SVM training method based on block chain

A training method, blockchain technology, applied in the field of artificial intelligence machine learning, to achieve the effect of ensuring high efficiency and security

Active Publication Date: 2020-05-05
BEIJING INSTITUTE OF TECHNOLOGYGY
View PDF6 Cites 6 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The third is to solve the problem of model training efficiency. Most of the model training work is carried out locally at

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
  • Safety SVM training method based on block chain
  • Safety SVM training method based on block chain
  • Safety SVM training method based on block chain

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0056] A blockchain-based secure SVM training method, such as figure 1 shown. figure 1 A safe SVM co-training scenario is described. There are 3 data providers participating in the training, namely |N|=3. Data providers are responsible for collecting, processing and building their own data sets. There are differences in the attributes contained in the datasets between different data providers, so after the three data providers share the data, a dataset with comprehensive attributes can be formed. The three data providers are also the trainers of the model. Under the condition that the original training data and intermediate calculation results are not leaked, they collaborate to complete the training of the SVM model based on the data sets with different attributes. During the training process, the blockchain-based data platform connects various data providers to provide a decentralized collaborative training environment.

[0057] During the model training process, the dat...

Embodiment 2

[0069] This embodiment is to compare the results of the scene after the number of data providers of the present invention is expanded from 3, to verify the change of the accuracy rate of the method adopted in the invention when the number of data providers is different. At the same time, increase the model training under the dataset Australian Credit Approval Data (ACAD). Assume that the numbers of data providers are 3, 4, and 5 respectively. That is, the data set is vertically cut into 3, 4, and 5 parts according to the attributes. Follow steps 1 to 4 to train the model, and count the classification accuracy of the model. Table 1 shows the statistical results of running time when the number of data providers is 3, and the accuracy of the model is shown in Table 2; when the number of data providers expands, the accuracy of the model is shown in Table 3.

[0070] Table 1 Running time statistics table

[0071]

[0072]

[0073] Table 2 Accuracy comparison result table ...

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 safety SVM training method based on a block chain, can effectively solve the problem of user data privacy leakage during machine learning model training under a block chain platform, and belongs to the technical field of artificial intelligence machine learning. According to the method, under the condition that a trusted third party is not introduced, a secure data sharing platform is established based on a block chain technology; when each data provider shares the calculation intermediate value, the calculation intermediate value to be shared is encrypted through a threshold homomorphic encryption algorithm and then is shared to the platform, so that the security of the user data in the sharing process is ensured; most of model training work is locally carried out on a data provider, and calculation input is based on plaintext data, so that the high efficiency of model training is ensured. The method is especially suitable for a scene oriented to a vertical cut data set multi-party cooperative training model.

Description

technical field [0001] The invention relates to a block chain-based secure SVM (Support Vector Machine, support vector machine) training method for multiple user data sets, belonging to the technical field of artificial intelligence machine learning. Background technique [0002] With the development and application of information technology in all walks of life, a large amount of data is generated every day, such as medical care, Internet of Vehicles, etc. As an effective data analysis method, machine learning is widely used in these scenarios. Among various machine learning methods, SVM (Support Vector Machine, Support Vector Machine) is a common and efficient method. For example, in Vehicular Social Networks (VSNs), SVM is used to train recommender systems. In medical scenarios, SVM is used for disease prediction. In addition to effective data analysis methods, the training set used for model training is closely related to the effect of the model. In the Internet of V...

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): G06K9/62G06F21/62G06F21/60
CPCG06F21/602G06F21/6245G06F18/2411G06F18/214
Inventor 沈蒙张杰唐湘云祝烈煌
Owner BEIJING INSTITUTE OF TECHNOLOGYGY
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