AI Model Retuning Framework for Wireless Network Reliability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
AI/ML models in wireless communication systems experience performance fluctuations due to deployment in different geographic locations, channel conditions, and over time, leading to degradation in user equipment (UE) performance.
Innovation Solution
A method is disclosed that involves evaluating the performance of a first AI/ML model used by UE in a wireless network, sending configuration to a network entity for model retuning, monitoring and evaluating the performance of a second retuned AI/ML model, and storing the retuned model for use by other UE.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single AI/ML model is deployed for all user equipment, then device complexity is reduced, but performance reliability deteriorates due to performance fluctuations in different geographic locations and channel conditions
Solution Approach 1:
The patent segments the single model deployment approach into multiple model versions (first AI/ML model and second AI/ML model) with different lineages. The network element maintains and manages multiple model versions to address performance fluctuations in different geographic locations and channel conditions, thereby improving reliability without significantly increasing complexity.
Solution Approach 2:
The patent changes the parameter of model versioning by introducing model lineage identification. The network element can identify and select appropriate model versions based on performance evaluations, allowing it to adapt to different deployment conditions while maintaining manageable complexity through systematic parameter tracking.
2Reliability
If AI/ML model performance is continuously monitored and retuned, then model reliability is improved, but network operation complexity increases
Solution Approach 1:
The patent implements a feedback mechanism where the network element continuously monitors performance of AI/ML models and triggers retuning operations based on performance evaluations. This feedback loop ensures reliability improvement while managing complexity through automated decision-making based on performance thresholds and model lineage tracking.
Solution Approach 2:
The patent performs preliminary actions by maintaining multiple model versions in advance (first and second AI/ML models with different lineages). When performance degradation is detected, the system can quickly switch to a pre-prepared retuned model version, reducing the complexity of real-time retuning operations while maintaining high reliability.
3Adaptability or versatility
If multiple AI/ML model versions are maintained for different lineages, then adaptability to different conditions is improved, but loss of information increases due to model management overhead
Solution Approach 1:
The patent creates a universal model management framework that handles multiple AI/ML model versions with different lineages through a single standardized process. The network element uses unified performance evaluation and retuning procedures that work across all model versions, improving adaptability to different conditions while minimizing information loss through consistent management practices.
Data Source
AI summary
A network element evaluates performance of a first AI/ML model being used by a UE. The network element sends, based on the evaluation, configuration to a network entity involved in performing model retuning of the first AI/ML model to aid in the model retuning. The network element monitors and evaluates performance of a second AI/ML model that is a retuned version of the first AI/ML model. The first and second AI/ML models are from a same lineage of AI/ML models. The network element stores, in response to the evaluation of the second AI/ML model, the second AI/ML model for use by other UE(s). A UE receives the configuration, and performs operation(s) to aid in the performing retuning. The retuning creates a second AI/ML model that is a retuned version of the first AI/ML model. The UE switches from the first AI/ML model to the second AI/ML model.