AI Model Update Signaling for Dynamic Wireless Communication
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 permission problems, which cannot be timely updated by the core network device, 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 improving transmission efficiency by negotiating and indicating progress, ensuring quick deployment of updated models.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If the core network device separately transmits AI models to terminal devices and access network devices, then the AI model can be deployed to multiple devices, but the core network device cannot perceive changes in the wireless environment and cannot update the AI model in a timely manner
Solution Approach 1:
The patent implements a feedback mechanism where terminal devices and access network devices report AI model performance metrics and environmental change information back to the core network device. This enables the core network device to perceive wireless environment changes and trigger timely AI model updates, resolving the contradiction between model adaptability and update timing.
Solution Approach 2:
The patent establishes preliminary action by having the core network device proactively push AI model updates to terminal devices and access network devices based on predicted environmental changes or performance degradation thresholds, rather than waiting for devices to request updates. This reduces the time delay in model updates while maintaining adaptability.
2Reliability
If the core network device transmits updated AI model information to the first 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 dividing updated model information into multiple parts or packets that can be transmitted incrementally. This allows the system to update AI models reliably while managing transmission resources efficiently, as resources can be allocated in manageable chunks rather than requiring large contiguous resource blocks.
Solution Approach 2:
The patent applies partial action by transmitting only the necessary portions of AI model updates rather than complete models every time. When only specific model parameters or layers need updating, only those partial updates are transmitted, reducing resource consumption while maintaining update reliability.
3Productivity
If the AI model is updated frequently to maintain performance in dynamic environments, then application effectiveness improves, but system complexity and overhead increase
Solution Approach 1:
The patent implements periodic action by establishing regular update intervals and scheduling mechanisms for AI model updates. Instead of continuous or ad-hoc updates, the system performs updates at predetermined intervals or when triggered by specific events, reducing negotiation and transmission complexity while maintaining application effectiveness through systematic update cycles.
Data Source
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.