ML Model Assurance for Wireless Device RAN Configuration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for managing wireless devices in communication networks struggle to ensure that devices are using the correct and current version of Machine Learning (ML) models, leading to suboptimal or erroneous performance in radio network operations.
Innovation Solution
A method where a Radio Access Network (RAN) node sends an ML model Assurance Information (MAI) Request to a wireless device, which responds with ML model characteristic information, allowing the RAN node to configure RAN operations accordingly and ensure the correct version of the ML model is used.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If the network downloads ML models to devices, then device-side inference capability is improved, but the network cannot verify whether devices are using the correct model version
Solution Approach 1:
The patent implements a feedback mechanism where the network requests ML model information from devices, devices report their model characteristics, and the network verifies model version accuracy. This closed-loop feedback system enables the network to monitor and ensure devices are using correct model versions while maintaining device-side inference autonomy.
Solution Approach 2:
The patent applies preliminary action by having the network send ML models to devices in advance before inference is needed. This allows devices to have the models readily available for execution, improving inference capability while the subsequent verification process ensures model version correctness.
2Loss of energy
If devices execute ML models locally, then base station resource usage is reduced, but device implementation may deviate from the original model
Solution Approach 1:
The verification mechanism provides feedback to detect when devices deviate from the original model implementation. By comparing reported model characteristics against expected values, the network can identify implementation deviations and take corrective action, ensuring model fidelity while maintaining local execution benefits.
Solution Approach 2:
The patent replaces direct mechanical control of model execution at the base station with a verification-based approach. Instead of centrally executing models and controlling device behavior, the system uses information exchange and verification mechanisms to ensure proper model usage, reducing base station resource consumption while maintaining implementation accuracy.
3Quantity of substance
If devices use old ML models, then device memory usage is reduced, but radio network operation performance deteriorates
Solution Approach 1:
The feedback mechanism enables the network to monitor which model versions devices are using and can request updates when performance deterioration is detected. This allows dynamic management of model versions based on actual performance needs rather than static memory constraints.
Solution Approach 2:
The patent introduces dynamics to the model management system, allowing model versions to be updated based on changing network conditions and performance requirements. Rather than devices statically holding old models to save memory, the system dynamically manages model versions to maintain optimal radio network operation performance.
Data Source
AI summary
A method (100) is disclosed for managing a wireless device that is operable to connect to a communication network, wherein the communication network comprises a Radio Access Network (RAN), and wherein the wireless device has available for execution a Machine Learning (ML) model that is operable to provide an output, on the basis of which a RAN operation performed by the wireless device may be configured. The method, performed by a RAN node of the communication network, comprises, on fulfilment of a trigger condition, causing an ML model Assurance Information, MAI, Request to be sent to the wireless device (110), the MAI Request comprising an indication of the ML model to which the MAI Request relates. The method further corpses receiving, from the wireless device, an MAI Response, wherein the MAI Response comprises ML model characteristic information generated by the wireless device using the ML model (120), and configuring the RAN operation performed by the wireless device according to the received MAI Response (130).