Efficient quality inspection model construction method and system for self-learning and updating training samples

A technology of training samples and construction methods, applied in neural learning methods, biological neural network models, neural architectures, etc., can solve problems such as long prediction time, prevent the generation of local optimal models, ensure efficiency, and improve recognition efficiency. Effect

Inactive Publication Date: 2019-02-19
心鉴智控(深圳)科技有限公司
View PDF0 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Similarly, the larger n, the better the effect, but also, the longer the prediction time

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
  • Efficient quality inspection model construction method and system for self-learning and updating training samples
  • Efficient quality inspection model construction method and system for self-learning and updating training samples
  • Efficient quality inspection model construction method and system for self-learning and updating training samples

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0049] It should be noted that, in the case of no conflict, the embodiments of the present invention and the features in the embodiments can be combined with each other.

[0050] Content implementation of the present invention is described as follows:

[0051] 1. The symbols involved in the algorithm are expressed as follows:

[0052] L: Labeled sample set;

[0053] U: no sample label;

[0054] Z k : proxy label of the kth sample set;

[0055] m 0 : parent model, W 0 : the weight of the network structure of the parent model;

[0056] m i : the i-th training sub-model, i=1,2...n; let n be 2 in this article, that is, there is a sub-model M 1 and M 2 ;

[0057] W i : the weight of the sub-model network structure;

[0058] g 1 (x) and g 2 (x): Two image enhancement functions.

[0059] Loss: loss function.

[0060] Second, the algorithm is explained in detail as follows:

[0061] first part:

[0062] 1) That is: first use the label data L to train the SqueezeNet...

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 method and a system for constructing an efficient quality inspection model of a self-learning and updating training sample, wherein two SqueezeNet Pro supervised learning models are adopted, fast k. The means algorithm acts as a preselected tag; The loss function of each model is optimized by comparing the two data enhancement algorithms. Conducting self-learning and training models; And update the weights of the parent model. The invention effectively improves the identification efficiency or the identification accuracy, improves the speed of the algorithm calculation, does not increase the time complexity of the operation, and can effectively prevent the occurrence of the local optimal situation.

Description

technical field [0001] The invention belongs to the field of quality inspection model algorithms, in particular to an efficient quality inspection model construction method and system for self-learning and updating training samples. Background technique [0002] For quality inspection tasks, the acquired data needs to be used to update the classification model algorithm in real time, so as to continuously improve the accuracy of the model. One of the most important difficulties encountered in this process is that the data obtained in real time has no labels, so the labels recognized by the model are simply used as training labels to further update the model. This is the most basic self-training method. It will make the model more convinced of what it thinks is right, thus increasing the over-fitting state of the model and failing to improve the accuracy of the model, and has no use value. [0003] In many application fields, semi-supervised learning (Semi-supervised) adopte...

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): G06K9/62G06N3/04G06N3/08
CPCG06N3/088G06N3/045G06F18/23213G06F18/214
Inventor 罗晓忠毛子靖林清华
Owner 心鉴智控(深圳)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