Network Handover Metadata Signaling for Machine Learning Models
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing communication systems lack efficient mechanisms for managing machine learning models during handover procedures between network apparatus and terminals, leading to potential disruptions and inefficiencies in training and execution of these models.
Innovation Solution
Implementing mechanisms for receiving, determining, and signaling metadata about machine learning models during handover procedures, allowing network apparatus and terminals to manage the execution and training of these models effectively, including the ability to signal requests and responses for further information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If machine learning models are transferred during handover without metadata management, then model continuity may be maintained, but network overhead and processing complexity increase significantly
Solution Approach 1:
The patent extracts only the essential metadata (model identifiers, execution status, training status) from the complete machine learning model data, transferring this compact representation during handover instead of the full models. This reduces network overhead while maintaining the ability to track and manage model continuity across network transitions.
Solution Approach 2:
The patent performs preliminary actions by pre-processing and categorizing machine learning model metadata before handover occurs. The network apparatus prepares model information in advance, organizing it into structured formats that facilitate efficient transfer and decision-making during the actual handover process, reducing processing complexity at the critical moment of transition.
2Loss of information
If all machine learning model metadata is transferred during handover, then complete model information is available, but data transmission load increases
Solution Approach 1:
The patent selectively extracts only the necessary metadata fields (model identifiers, execution status, training status) from the complete machine learning model dataset, transferring this minimized information set during handover. This extraction approach maintains sufficient model information for continuity while dramatically reducing the data transmission load compared to transferring complete models or all possible metadata.
Solution Approach 2:
The patent applies different levels of information detail to different metadata elements, transferring essential identification and status information during handover while leaving more detailed model parameters to be retrieved on-demand or maintained locally. This local quality differentiation optimizes the balance between information completeness and transmission efficiency.
3Adaptability or versatility
If machine learning model execution decisions are made without standardized signaling, then flexibility is maintained, but communication efficiency between network apparatus and terminal decreases
Solution Approach 1:
The patent creates a universal signaling framework using standardized message types (RRC messages, system information blocks) that can carry machine learning model metadata and execution decisions. These existing communication protocols are extended to serve multiple functions including model information transfer, execution status notification, and training status reporting, improving communication efficiency while maintaining flexibility through configurable message content.
Data Source
AI summary
There is provided a network apparatus that is caused to receive as part of a handover procedure for handover of a terminal to the network apparatus, metadata about at least one machine learning model accessible for execution and/or training by the terminal, determining whether or not the terminal should execute and/or train the at least one machine learning model after the terminal is handed over to the network apparatus; and signal the result of the determining to the terminal.