Course label generation method, apparatus and device, and medium

Abstract

The invention relates to the field of artificial intelligence, and provides a course label generation method, apparatus and device, and a medium, which can more accurately extract keywords based on content understanding, indirectly improve the reasonability of label generation, extract time sequence features by using an LSTM network, enable the extracted features to be more matched with the attributes of course data to be processed, and improve the accuracy of course label generation. According to the method, the text is subjected to text recognition through the CNN-Attention network, recognition of important characters can be more focused, semantics are deeply understood through regression analysis, paragraph information of each text is better expressed, the accuracy of text representation is improved, more accurate keyword extraction is achieved, and through further cleaning and correction processing. Therefore, the generated label set is more accurate, and automatic generation of course labels is realized. The invention also relates to a blockchain technology, and the involved model can be stored in the blockchain.

Description

technical field [0001] The present invention relates to the technical field of artificial intelligence, in particular to a method, device, equipment and medium for generating course labels. Background technique [0002] Online education and training has become an important means of Internet education. Online education is mainly carried out through live video, course recording, and playback. The label information of live courses is very important and meaningful for specific tasks such as course classification, course retrieval and course label generation. [0003] The current labeling methods mainly include manual labeling and machine automatic labeling. For the method of manual labeling, the creator generally performs labeling at the same time as the course is generated; for the method of machine labeling, it mainly infers the text through the course title, content summary, etc., and then performs labeling according to the inferred results. mark. [0004] However, the man...

AI Technical Summary

Problems solved by technology

[0004] However, the manual labeling method is highly subjective, not only time-consuming, but also low in accuracy; the current machine automatic labeling method is not very accurate, resulting in a serious lack of course labels or title content. The label of the course itself is inconsistent with the content of the course itself

Images