Corpus data labeling method based on swarm intelligence
What is Al technical title?
Al technical title is built by PatSnap Al team. It summarizes the technical point description of the patent document.
A corpus and data technology, applied in the field of corpus labeling, to reduce the error rate of labeling, shorten the development cycle and labeling cycle, and improve the accuracy rate
Inactive Publication Date: 2018-11-23
SHENZHEN QIANHAI YYD ROBOT CO LTD
View PDF3 Cites 4 Cited by
Summary
Abstract
Description
Claims
Application Information
AI Technical Summary
This helps you quickly interpret patents by identifying the three key elements:
Problems solved by technology
Method used
Benefits of technology
Problems solved by technology
[0004] In order to solve the shortcomings of traditional labeling methods, the 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
Click on the blue label to locate the original text in one second.
Reading with bidirectional positioning of images and text.
Smart Image
Examples
Experimental program
Comparison scheme
Effect test
Example Embodiment
[0034] Embodiment one:
[0035] refer to figure 1 As shown, the annotation system model is constructed according to the corpus data annotation process. The annotation system model mainly includes four parts: annotation model 1, annotation data generator 2, data to be annotated 3, and annotated data 4.
[0036] refer to figure 2 , 3 , 4, and 5, on the basis of the labeling system model, a corpus data labeling method based on crowd intelligence is proposed, including the following steps:
[0037] S1. At the initial stage of labeling, the user enters the data to be labeled 3, and initializes the labeling model 1 and the labeling data generator 2 at the same time;
[0038] S2. The annotation data generator 2 extracts a certain amount of data from the data to be annotated 3 and the annotated data 4 to generate annotated data 7 for the annotator;
[0039] S3. The labeling model 1 performs iterative training based on the labeled data 4, and then automatically labels the labeling...
Example Embodiment
[0043] Embodiment two:
[0044] refer to Image 6 , 7 , 8, a corpus data tagging system based on crowd intelligence based on the method of Embodiment 1, the system includes a Web background server 100, a GPU server 200, and multiple computers 300 equipped with browsers. Web background server comprises labeling data generating unit 101, labeling data storage unit 102, instruction sending unit 103; Many computers 300 equipped with browsers are used for labeling personnel to log in data labeling system, input data 3 to be labelled, and label data 7 For labeling or confirmation, the GPU server 200 is used to run the labeling model unit 201, and perform iterative training according to the marked data 4 and instructions provided by the web background server 100, wherein the labeling model unit 201 is used to mark the labeling data 7, and in The R&D process generates annotated models1.
[0045]The input of the tagging model unit 201 is a sentence or segment in the corpus data to b...
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
Login to view more
Abstract
The invention discloses a corpus data labeling method based on swarm intelligence. A user inputs to-be-labeled data; a labeling model and a labeled data generator are initialized; the labeled data generator extracts a certain number of data from the to-be-labeled data and labeled data for labeling personnel to generate the labeled data; the labeling model is generated in a research and developmentprocess; the labeling model performs iterative training according to the labeled data, and then performs automatic labeling on the to-be-labeled data to generate automatically labeled data of the labeling model; and the labeling personnel perform labeling or confirmation on the labeled data, and store the labeled data. Meanwhile, based on the data labeling method, the invention provides a corpusdata labeling system based on the swarm intelligence. According to the labeling method and system provided by the invention, the development process and the labeling process can be combined, so that the research and development period and the labeling period can be remarkably shortened, and the accuracy of data labeling is improved.
Description
technical field [0001] The patent of the present invention relates to the field of corpus labeling, in particular to a corpus data labeling method and system based on crowd intelligence. Background technique [0002] Today, with the rapid development of the Internet and artificial intelligence, the need for data labeling is becoming more and more urgent. In the face of data labeling requirements, the existing labeling methods use word, excel and other text forms, or develop a special web page system for organizing labeling corpus. Annotators then annotate the corpus in the form of text or web pages. [0003] The problem with the existing labeling methods is that the labelers do their own thing and label their own data, and there are also differences in labeling standards between the labelers; the labeling and R&D processes are independent of each other and cannot be developed collaboratively; Correction, or can only be corrected by secondary marking. The above problems li...
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
Application Date:The date an application was filed.
Publication Date:The date a patent or application was officially published.
First Publication Date:The earliest publication date of a patent with the same application number.
Issue Date:Publication date of the patent grant document.
PCT Entry Date:The Entry date of PCT National Phase.
Estimated Expiry Date:The statutory expiry date of a patent right according to the Patent Law, and it is the longest term of protection that the patent right can achieve without the termination of the patent right due to other reasons(Term extension factor has been taken into account ).
Invalid Date:Actual expiry date is based on effective date or publication date of legal transaction data of invalid patent.
Login to view more 
Login to view more