Terminal Recommendation Architecture Using a Shared Algorithm Library

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing mobile terminals face code redundancy issues due to the inclusion of machine learning algorithms in the service code of each intelligent service, leading to inefficiencies.

Innovation Solution

A method and terminal implementation that utilize a local algorithm library, allowing the terminal to invoke algorithms based on configuration parameters and user data, reducing the need for algorithm duplication in individual service codes.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If machine learning algorithms are carried in service code of each intelligent service, then intelligent recommendation services can be implemented on the mobile terminal, but code redundancy is caused

Engineering Contradiction:
Improveintelligent recommendation service capabilityVSAvoidcode redundancy
Core Design Contradiction:
Adaptability or versatilityVSQuantity of substance

Solution Approach 1:

The patent merges the machine learning algorithms from individual service codes into a shared algorithm library. The terminal obtains algorithms from this common library rather than embedding them in each service's code, thereby eliminating redundancy while preserving intelligent recommendation capabilities across multiple services.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The algorithm library serves as a universal resource that multiple intelligent services can access and utilize. By making the algorithm library multi-functional and shared across different services, the system avoids duplicating the same algorithms in each service code while maintaining adaptability and versatility of the recommendation system.

Inventive Principle:
Principle #6Universality (Multi-functionality)

2Reliability

If machine learning algorithms are embedded in each service code, then each service can independently perform user preference prediction, but computational efficiency and resource utilization deteriorate

Engineering Contradiction:
Improveindependent prediction capabilityVSAvoidcomputational efficiency
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The patent combines the computational resources and algorithms into a shared library that multiple services access collectively. This merging approach maintains the ability to perform user preference prediction while improving computational efficiency by avoiding redundant algorithm execution and reducing overall resource consumption.

Inventive Principle:
Principle #5Merging (Combining)

Data Source

PatentEP3553676B1Smart recommendation method and terminal
Publication Date: 2025.12.31 HUAWEI TECH CO LTD
  • EP3553676B1 patent drawingFigure 1
  • EP3553676B1 patent drawingFigure 2
  • EP3553676B1 patent drawingFigure 3

AI summary

Embodiments of the present invention provide an intelligent recommendation method and a terminal. The method includes: receiving, by a terminal, a service request of a recommendation application, where the service request includes a configuration file, the configuration file includes at least an algorithm parameter and portfolio information that is used to identify an algorithm portfolio structure, and the recommendation application is included in the terminal; invoking, by the terminal, at least one algorithm from an algorithm library of the terminal based on the algorithm parameter and the portfolio information; predicting, by the terminal based on the at least one algorithm and user data of the terminal, at least one service currently required by a user of the terminal, where the user data of the terminal is stored in a personal database that is located in the terminal; and displaying, by the terminal, at least one icon corresponding to the at least one service. Code redundancy can be reduced in the embodiments of the present invention.