Computing device and method of operation thereof
By utilizing category-specific content consumption information from computing devices, heterogeneous content similarity can be directly extracted from the user's multi-device consumption history. This solves the problems of data quality dependence and computational complexity in heterogeneous content recommendation in existing technologies, and achieves fast and accurate recommendation results.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- SAMSUNG ELECTRONICS CO LTD
- Filing Date
- 2024-10-14
- Publication Date
- 2026-06-02
AI Technical Summary
Existing collaborative filtering recommendation systems suffer from problems such as high dependence on data quality, high computational complexity, need for active human intervention, and low performance when data is insufficient in heterogeneous content recommendation.
By computing device-specific content consumption information based on categories, user-specific heterogeneous content consumption information is obtained. Heterogeneous content is recommended using a similarity matrix, avoiding reliance on complex neural networks and directly extracting similarity from the user's multi-device consumption history.
It enables fast and accurate content recommendation in heterogeneous content recommendation, reduces computational complexity and dependence on metadata, and improves the autonomy and efficiency of recommendation.
Smart Images

Figure CN122139202A_ABST