UE-Initiated Two-Sided AI Model Updates for Consistent Inference
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Two-sided AI/ML models in communication networks face challenges in maintaining consistent model inference across nodes due to data set changes, leading to potential connection failures and performance issues during updates.
Innovation Solution
A UE-initiated model-update mechanism is implemented, allowing the user equipment to trigger an uplink transmission with specific parameters for updating the ML model when data sets change, ensuring consistent model inference without sacrificing performance or latency by coordinating with the network.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the ML model is updated at the UE side when data sets change, then the model inference consistency is improved, but the connection failures and performance issues occur during updates
Solution Approach 1:
The network configures the UE with update parameters in advance through RRC signaling, including uplink resources, format, and information elements. This preliminary configuration enables the UE to perform model updates proactively when data sets change, ensuring inference consistency while preventing connection failures through pre-established update mechanisms.
Solution Approach 2:
The mechanism implements feedback through network configuration and UE reporting. The network provides update parameters via RRC signaling, and the UE uses these parameters to trigger uplink transmissions for model updates. This feedback loop ensures that model updates are coordinated between network and UE, maintaining consistency while preventing harmful connection failures.
2Loss of time
If the UE triggers uplink transmission for model updates, then the model update timing is improved, but the network coordination complexity increases
Solution Approach 1:
The invention changes the state of network configuration parameters by introducing specific RRC signaling elements that enable UE-initiated updates. The network configures update parameters including uplink resource allocations, transmission formats, and information elements that the UE uses to trigger updates. This parameter-based approach improves update timing while managing coordination complexity through standardized signaling procedures.
3Manufacturing precision
If the model update parameters are configured via RRC signaling, then the update configuration precision is improved, but the signaling overhead increases
Solution Approach 1:
The invention extracts and separates the model update configuration into specific RRC signaling elements and information elements. By taking out the update parameters (uplink resources, format, triggering conditions) as distinct configurable elements, the system achieves precise update configuration while managing signaling overhead through structured, modular information elements that can be efficiently transmitted and processed.
Data Source
AI summary
A method comprising obtaining, at a terminal assigned to a network, a configuration with parameters that enable the terminal to initiate an uplink, UL, transmission related to an update at the terminal at least for a terminal part of a two-sided model used at the terminal, wherein the two-sided model is used for joint model inference at the terminal side and the network side; determining that an update of the two-sided model is required; initiating, based on the obtained configuration, a first UL transmission; transmitting the initiated first UL transmission indicating the required update; receiving a response comprising update information related to the required update; and based on the received response, establishing an update of the terminal part of the two-sided model.