AI Model Accuracy Signaling for Reliable Network Inference
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing communication networks face challenges in ensuring the reliability and accuracy of artificial intelligence (AI) model inference results due to deviations in model accuracy information, which can affect the usability of inference outcomes.
Innovation Solution
A method and apparatus for information transmission that includes sending and receiving accuracy information of AI models between devices, allowing devices to trigger model execution and generate models, thereby enabling devices to understand the data used to obtain accuracy information and perform subsequent processing based on this information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If AI model inference is performed in communication networks, then intelligent data analysis capability is improved, but reliability of inference results deteriorates due to accuracy deviations
Solution Approach 1:
The patent implements a feedback mechanism where accuracy information of AI models is collected from multiple sources (third devices that generate models and second devices that trigger execution), evaluated by the first device, and used to determine whether to perform inference. This feedback loop ensures that inference reliability is maintained by continuously monitoring and assessing model accuracy before deployment.
2Reliability
If accuracy information is transmitted between devices, then reliability of inference is improved, but communication overhead increases
Solution Approach 1:
The patent extracts and transmits only the essential accuracy information needed for reliability assessment, rather than transmitting complete model datasets or raw inference results. This selective extraction of critical information (accuracy metrics, model performance data) reduces communication overhead while maintaining the ability to assess inference reliability.
3Reliability
If multiple devices are involved in model execution and accuracy assessment, then reliability of accuracy information is improved, but device complexity increases
Solution Approach 1:
The patent designs a multi-functional architecture where the first device performs both inference execution and accuracy evaluation, the second device handles model triggering and accuracy reporting, and the third device generates models and provides accuracy information. This universal design allows each device to serve multiple purposes, improving reliability through distributed assessment while managing complexity through role specialization.
Data Source
AI summary
This application discloses an information transmission method and apparatus and a device, and pertains to the field of communication technologies. The method according to an embodiment of this application includes: sending, by a first device, first information corresponding to first accuracy information of a first model to a second device or a third device, where the first information is used to describe information used to obtain the first accuracy information; and the second device is a device that triggers the first device to execute the first model, and the third device is a device that generates the first model.