Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Transfer learning algorithm based on active learning

A technology of transfer learning and active learning, applied in the field of machine learning, it can solve the problem of high cost of data labeling, and achieve the effect of performance improvement and good transfer ability.

Pending Publication Date: 2020-05-22
NANJING UNIV OF AERONAUTICS & ASTRONAUTICS
View PDF0 Cites 6 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, behind the high accuracy rate is a large amount of high-quality labeling data. The reality is that the cost of data labeling is extremely expensive, which is unaffordable for some small companies and even some large companies.

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
  • Transfer learning algorithm based on active learning
  • Transfer learning algorithm based on active learning
  • Transfer learning algorithm based on active learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0028] A migration learning algorithm based on active learning proposed by the present invention will be described in detail below with reference to the accompanying drawings.

[0029] like figure 1 As shown, the migration learning algorithm based on active learning proposed in the present invention comprises the following steps:

[0030] Step 1) Determine the input variables of the algorithm, including the source domain and target domain data sets to be trained, the current iteration number t, the current model Mt, and the selected marked data set Q;

[0031] Step 2) Use an unsupervised domain adaptive algorithm to train the source and target domain datasets to obtain an initialized model M0;

[0032] Step 3) For each sample point x in the target domain data set X, calculate its feature x'=conv(x) extracted after passing through the convolutional layer;

[0033] Step 4) Calculate feature discriminative index

[0034] Step 5) Calculate the uncertainty index of sample x, w...

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 transfer learning algorithm based on active learning, and belongs to the field of machine learning. For a general unsupervised transfer learning algorithm, a large number ofresearches exist at present, but on this basis, the improvement of the algorithm performance in a target field can be obtained at a relatively low sample labeling cost. The active transfer learning algorithm accesses a batch of data based on an active sampling method to finely tune and update network parameters after an unsupervised domain self-adaption process is carried out, so that the extracted features have good migration capability and good discrimination capability. In the invention, an active sampling strategy is not only based on a traditional information entropy method, but also provides an evaluation index of one characteristic under a transfer learning background.

Description

technical field [0001] The invention relates to the technical field of machine learning, in particular to a migration learning algorithm based on active query. Background technique [0002] In recent years, deep learning has achieved great success, such as computer vision, speech recognition, natural language processing, etc. State-of-the-art results have been achieved on several standard datasets. However, behind the high accuracy rate is a large amount of high-quality labeling data. The reality is that the cost of data labeling is extremely expensive, which is unaffordable for some small companies and even some large companies. And the real intelligent technology requires the ability to draw inferences from one instance, which means that the learned model can be transferred in similar scenarios, instead of training from scratch for each task. Based on the above requirements, transfer learning has gained more and more attention. [0003] The key to transfer learning tech...

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): G06N3/08G06K9/62
CPCG06N3/088G06F18/217
Inventor 关东海张琦
Owner NANJING UNIV OF AERONAUTICS & ASTRONAUTICS
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products