Migration knowledge determination method, device and equipment and readable storage medium

A technology for determining a method and knowledge, applied in the transfer knowledge determination device, equipment and readable storage medium, in the field of transfer knowledge determination, and can solve the problems of inability to use the learner, not sharing, etc.

Active Publication Date: 2019-05-03
BEIJING SANKUAI ONLINE TECH CO LTD
View PDF9 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] In the prior art, the transfer knowledge determination method is usually limited to one or a few learners, the transfer depends on the ability of the learner, it cannot be used for any learner, and it is not a sharing in the true sense.

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
  • Migration knowledge determination method, device and equipment and readable storage medium
  • Migration knowledge determination method, device and equipment and readable storage medium
  • Migration knowledge determination method, device and equipment and readable storage medium

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0030] refer to figure 1 , which shows a flow chart of steps of an embodiment of a migration knowledge determination method embodiment of the present invention, which may specifically include the following steps:

[0031] Step 101, obtaining at least two text raw data based on different classification tasks;

[0032] In the embodiment of the present invention, in the technical field of machine learning, in order to obtain enough features for the learner to learn, it is usually necessary to collect sufficient data for the learning task or learning field. In the case of insufficient data collection, it is necessary to migrate other data from different fields or the same type of tasks or with similar characteristics to the current task to meet the training data requirements of the current machine learning task.

[0033] As mentioned above, two or more raw text data for different classification tasks are collected. Usually, raw text data refers to raw data without any data proces...

Embodiment 2

[0049] refer to figure 2 , which shows a flow chart of steps of an embodiment of a migration knowledge determination method embodiment of the present invention, which may specifically include the following steps:

[0050] Step 201, acquiring at least two text raw data based on different classification tasks;

[0051] This step is the same as step 101, and will not be described above here.

[0052] Step 202, performing word segmentation on the text raw data based on different classification tasks to obtain word segmentation data;

[0053] Step 203, converting the word segmentation data into feature vectors of preset dimensions;

[0054] Specifically, firstly, after obtaining raw text data based on different fields, use a word segmentation tool to segment the raw data; and then further convert the raw text data after word segmentation into feature vectors. Among them, matching word segmentation tools can be used according to different word segmentation purposes to achieve wo...

Embodiment 3

[0098] refer to image 3 , shows a flow chart of the steps of an embodiment of a method for determining the minimum contained ball sharing data of the present invention, which may specifically include the following steps:

[0099] Step 301, acquiring at least two data sets based on preset classification tasks;

[0100] In this example, if the model is to be trained for a certain classification task / domain 1, and the marked data in the acquisition task / domain 1 is insufficient, then there are part of the same classification task / domain 1 in classification tasks / domain 2 and 3 resource data, then obtain the corresponding raw data of classification tasks / domains 2 and 3, and perform preprocessing and classification to generate different data sets.

[0101] Of course, the number of tasks / fields is not limited, and data sets of multiple classification tasks / fields may be obtained for enrichment of data features, which is not limited in this embodiment.

[0102] Step 302, obtainin...

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 migration knowledge determination method, devcie, equipment, and a readable storage medium. The method comprises the steps of obtaining at least two pieces of text generationdata based on different classification tasks; mapping the text generation data based on the different classification tasks to a high-dimensional vector space of a preset dimension; obtaining correlation data of the text generation data based on the different classification tasks in the high-dimensional vector space; and mining shared data in the correlation data, and determining the shared data asmigration knowledge. The problem that knowledge migration is completed by depending on a learner in the knowledge migration field in the prior art is solved.

Description

technical field [0001] The present invention relates to the technical field of machine learning, in particular to a transfer knowledge determination method, a transfer knowledge determination device, equipment and a readable storage medium. Background technique [0002] In the field of machine learning technology, especially in supervised learning methods, it is necessary to have sufficient labeled training data in a field in order to obtain enough data features through machine learning and make predictions on target data. However, in In practical applications, it is often encountered that the training data is insufficient, for example, there is a classification task A in one domain, but there is only enough training data for the classification task B in another domain, where the data may have different data distributions , in this regard, it is usually necessary to migrate the data of classification task B to classification task A, that is, to share data in different fields...

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): G06K9/62G06F16/35
Inventor 刘凡
Owner BEIJING SANKUAI ONLINE 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