Communication Device AI Model Selection by Local Conditions

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing communication systems face challenges in efficiently utilizing AI models for device collaboration due to high overheads and inefficiencies in selecting and switching AI models based on device conditions.

Innovation Solution

A communication method that allows devices to select and switch AI models based on predefined conditions, reducing overheads by using correspondence information exchanged between devices.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If the second device selects AI model for the first device based on measurement information, then the AI model can be properly selected, but the overheads of obtaining measurement information increase

Engineering Contradiction:
ImproveAI model selection accuracyVSAvoidoverheads of obtaining measurement information
Core Design Contradiction:
Measurement precisionVSLoss of information

Solution Approach 1:

The first device autonomously selects the AI model based on its own measurement information and the correspondence provided by the second device, eliminating the need for the second device to obtain and process measurement information. This self-service approach reduces overheads while maintaining selection accuracy.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent extracts the AI model selection function from the second device and transfers it to the first device. The second device only provides the correspondence between conditions and AI models, while the actual selection is performed by the first device using its local measurement information, reducing the information overhead required.

Inventive Principle:
Principle #2Taking out (Extraction)

2Loss of information

If the first device autonomously selects AI model based on measurement information, then the overheads of obtaining measurement information are reduced, but the efficiency of determining and using AI model may be insufficient

Engineering Contradiction:
Improveoverheads of obtaining measurement informationVSAvoidefficiency of determining and using AI model
Core Design Contradiction:
Loss of informationVSProductivity

Solution Approach 1:

The second device pre-provides the correspondence between conditions and AI models to the first device before the actual selection is needed. This preliminary action enables the first device to efficiently select the appropriate AI model by simply matching its current measurement information against the pre-provided correspondence, improving determination efficiency.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The first device autonomously performs the AI model selection process using its own measurement information and the pre-provided correspondence, without requiring the second device to intervene in the selection process. This self-service mechanism improves efficiency by eliminating communication overheads during the selection phase.

Inventive Principle:
Principle #25Self-service

3Adaptability or versatility

If multiple AI models are supported with different conditions, then the adaptability of AI model selection is improved, but the device complexity increases

Engineering Contradiction:
ImproveAI model selection adaptabilityVSAvoidcomplexity of AI model management
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent segments the AI model management by dividing the correspondence information into multiple independent condition-AI model pairs. Each condition (signal quality, application scenario, resource status, location, capability, or performance) is evaluated independently, and the corresponding AI model is selected based on the first matching condition, simplifying the management complexity while maintaining high adaptability.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies different AI models based on local conditions specific to the first device's current state. Each condition represents a different aspect of the device's local environment or state, and the corresponding AI model is optimized for that specific condition, enabling adaptive selection without requiring complex global optimization.

Inventive Principle:
Principle #3Local quality

Data Source

PatentUS20250274354A1Communication method and apparatus
Publication Date: 2025.08.28 HUAWEI TECH CO LTD
  • US20250274354A1 patent drawing
  • US20250274354A1 patent drawing
  • US20250274354A1 patent drawing

AI summary

This application discloses a communication method and apparatus. The method includes: A first device may receive first information from a second device. The first information may indicate a correspondence between at least one condition and at least one AI model. When the first device meets a first condition, the first device may use a first AI model corresponding to the first condition, where the first condition is any one of the at least one condition. According to the method, the first device may select a to-be-used AI model based on the correspondence between the at least one condition and the at least one AI model indicated by the second device, so that efficiency of determining and using the AI model by the first device can be improved, and the first device can properly and effectively use the AI model, to improve efficiency of collaboration between the devices.