A method, apparatus, electronic device, and storage medium for recommending multimedia content.

By acquiring the content and object attributes of multimedia content, predicting the target display format, and filtering the feedback probability, the problem of insufficient personalization in multimedia content advertising exposure is solved, achieving more accurate personalized recommendations and improving advertising display effects.

CN117010967BActive Publication Date: 2026-06-30TENCENT TECH SHANGHAI
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
TENCENT TECH SHANGHAI
Filing Date
2022-10-19
Publication Date
2026-06-30

AI Technical Summary

Technical Problem

In existing technologies, the exposure of multimedia content advertisements lacks personalization. Uniform scene control leads to insufficient personalization and the inability to change the form according to the differences of different objects, resulting in model jitter and deviation.

Method used

By acquiring the content attributes, initial display format, and object attributes of multimedia content and the objects to be recommended, the target display format is predicted, and personalized multimedia content is selected for recommendation based on feedback probability. Personalized recommendations are then made using a content ranking model and a feedback probability model.

Benefits of technology

It enables personalized recommendations for multimedia content, improves advertising display effectiveness, reduces model jitter and bias, and enhances recommendation accuracy.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117010967B_ABST
    Figure CN117010967B_ABST
Patent Text Reader

Abstract

This application relates to the field of computer technology, and more particularly to a method, apparatus, electronic device, and storage medium for recommending multimedia content, to improve the accuracy of personalized multimedia content recommendations. The method includes: acquiring multiple multimedia content items to be recommended and recommended objects; predicting the target display format of each multimedia content item based on its content attributes and initial display format, and the object attributes of the recommended object; predicting the feedback probability of the recommended object for each multimedia content item based on its content attributes, target display format, and placement environment attributes, and the object attributes; selecting target multimedia content from the multimedia content items based on the obtained feedback probabilities, and recommending the target multimedia content to the recommended object based on the corresponding target display format. Because this application incorporates display format prediction of feedback probability, the accuracy of personalized multimedia content recommendations is improved.
Need to check novelty before this filing date? Find Prior Art