Base Station AI/ML Drift Monitoring Through User Grouping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing wireless communication networks face challenges in managing artificial intelligence and machine learning (AI/ML) model drift, leading to performance degradation due to changes in statistical properties of local data, which results in inefficiencies and increased signaling overhead.
Innovation Solution
Implementing user equipment grouping based on dataset category and lifecycle temporal range, with multi-threshold levels to prioritize communication resources and trigger mode switching for AI/ML models, using a lifecycle profile map for continuous monitoring and updating.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If AI/ML models are deployed in wireless communication networks, then communication efficiency and resource allocation improve, but model drift occurs due to changes in statistical properties of local data leading to performance degradation
Solution Approach 1:
The patent implements a feedback mechanism where the network device receives feedback information from user equipment regarding model drift detection results. Based on this feedback, the network device sends updated configuration information to adjust the AI/ML model parameters, creating a closed-loop system that continuously monitors and corrects model drift to maintain both communication efficiency and model accuracy.
Solution Approach 2:
The patent enables dynamic adaptation of AI/ML models by allowing model parameters to be adjusted in real-time based on detected drift conditions. The system transitions from static model deployment to dynamic model tuning, where configuration information is updated according to changing statistical properties of local data, thus maintaining model reliability while preserving productivity gains.
2Reliability
If AI/ML model monitoring and updating is implemented, then model accuracy is maintained, but signaling overhead increases
Solution Approach 1:
The patent implements partial monitoring and updating by selectively applying drift detection and model updates only when necessary. The system uses drift detection thresholds to determine when intervention is needed, avoiding continuous full-scale model updates. This partial action approach maintains model accuracy while significantly reducing the signaling overhead associated with constant model management communications.
3Productivity
If user equipment is grouped by dataset category and lifecycle temporal range, then resource allocation efficiency improves, but system complexity increases
Solution Approach 1:
The patent segments user equipment into distinct groups based on dataset category and lifecycle temporal range characteristics. This segmentation allows the network to apply different AI/ML model configurations and drift detection parameters to different UE groups, improving resource allocation efficiency by tailoring resources to specific device characteristics while managing complexity through structured categorization.
Data Source
AI summary
The present disclosure relates to a base station, a terminal and amethod for machine learning related communication between a network device and user equipments. The user equipments are grouped into at least one of several user groups, for handling of a detected machine learning model drift. If a machine learning model drift is detected,a machine learning model drift handling operation mode is selected for the user equipments of a user group.