Core Network AI Model Update Signaling Under Wireless Mobility
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
AI models in communication systems deteriorate due to mobility and time-varying wireless environments, leading to performance issues and area permissions, which the core network cannot timely address, affecting their application effectiveness.
Innovation Solution
A communication method where devices request updates to the core network with specific requirements, facilitating timely AI model updates and optimizing transmission efficiency by utilizing identifiers and indication information to manage AI model information exchange.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the core network device transmits the AI model to the terminal device and access network device, then the AI model can be deployed for use, but the AI model cannot be updated in a timely manner due to terminal device mobility and time-varying wireless environment changes
Solution Approach 1:
The patent implements a feedback mechanism where the terminal device and access network device monitor AI model performance and send update requests to the core network device when performance deterioration is detected. This feedback loop enables the core network device to receive real-time information about model effectiveness and initiate updates promptly, resolving the contradiction between maintaining model reliability and avoiding update delays.
Solution Approach 2:
The patent establishes preliminary action by having the core network device prepare updated AI models in advance and store them ready for transmission. When update requests are received from terminal or access network devices, the updated models are already prepared and can be transmitted immediately, eliminating the delay that would occur if models needed to be generated or retrieved during the update process.
2Loss of time
If the core network device transmits updated AI model information to the terminal device, then the AI model can be updated timely, but transmission resources may be insufficient or inefficient
Solution Approach 1:
The patent segments the AI model update transmission process by allowing the access network device to assist in forwarding update information from the core network to the terminal device. This segmentation of the transmission path enables more efficient utilization of available network resources, distributing the transmission load across multiple nodes rather than requiring a single direct transmission from the core network to the terminal.
Solution Approach 2:
The access network device serves as an intermediary in the AI model update transmission process. It receives update information from the core network device and forwards it to the terminal device, acting as a mediator that optimizes resource utilization by leveraging existing network infrastructure and reducing the burden on core network transmission resources.
3Reliability
If the AI model is updated frequently to maintain performance, then application effectiveness improves, but system complexity and overhead increase
Solution Approach 1:
The patent implements self-service by enabling the terminal device and access network device to autonomously monitor AI model performance metrics and automatically generate update requests when performance thresholds are violated. This eliminates the need for complex centralized monitoring and management systems, as the devices themselves perform the assessment and initiation of updates, reducing overall system complexity while maintaining performance.
Solution Approach 2:
The patent uses parameter changes by establishing performance thresholds and metrics that trigger update requests. When monitored parameters (such as model accuracy or performance metrics) change beyond acceptable thresholds, update requests are automatically generated. This parameter-based approach simplifies update management by providing clear, objective criteria for when updates are needed, reducing the complexity of decision-making in the update process.
Data Source
Figure 1~2
Figure 3~4
Figure 5~6
AI summary
Embodiments of this application provide a communication method and apparatus. The method may be used for updating an AI model. In the method, when an AI model currently used by a first device and a second device needs to be updated, the first device may send, to a core network device, requirement information for updating the AI model, and the core network device may determine updated AI model information based on the requirement information proposed by the first device. In this way, the first device and the second device can update the AI model based on the updated AI model information. This improves application effect of the AI model.