RAN Output Feedback Identifiers for Distributed ML Data Association
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current technologies face challenges in efficiently associating and managing feedback from multiple AI/ML model nodes, leading to difficulties in identifying which UE performance feedback corresponds to which AI/ML-based action, and lack of implementation-agnostic metrics for model performance evaluation.
Innovation Solution
A method for network nodes to provide an output feedback identifier that uniquely identifies AI/ML model outputs, enabling efficient collection and association of feedback data across nodes, allowing for improved model performance assessment and reduced signaling overhead.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If feedback data is collected from multiple AI/ML model nodes without unique identifiers, then data collection coverage is improved, but data association accuracy deteriorates
Solution Approach 1:
The patent segments the feedback data collection process by introducing unique output feedback identifiers that distinguish individual AI/ML model outputs. Each node receives configured identifier patterns, generates specific identifiers for its outputs, and associates these identifiers with collected feedback data. This segmentation enables precise tracking of which feedback corresponds to which model output across multiple distributed nodes.
2Measurement precision
If detailed feedback data is collected from all nodes, then model performance assessment accuracy is improved, but signaling overhead increases
Solution Approach 1:
The patent extracts the essential identification function from the feedback data itself by introducing separate output feedback identifiers. These identifiers are configured once at the OAM and reused across multiple feedback reporting instances. By separating the identification mechanism from the performance metric data, the system achieves accurate model performance assessment without proportionally increasing signaling overhead, as the identifier component remains constant while only the performance metrics need frequent reporting.
3Productivity
If multiple nodes report feedback independently without coordination, then data collection efficiency is improved, but data consistency deteriorates
Solution Approach 1:
The patent implements a universal identifier system where the OAM configures output feedback identifier patterns that all RAN nodes use consistently. This universal approach enables multiple nodes to independently collect and report feedback data simultaneously while maintaining consistency through the common identifier framework. The identifier patterns serve multiple functions: unique identification, correlation across nodes, and consistent data association throughout the distributed system.
Data Source
AI summary
In one aspect, a computer-implemented method performed by a first network node in a radio access network (RAN) is provided. The method includes obtaining an output from a machine learning (ML) model. The method includes obtaining an output feedback identifier for the output, wherein the output feedback identifier uniquely identifies the output. The method includes generating a first message, wherein the first message comprises the output feedback identifier. The method includes transmitting, towards a third network node, the first message comprising the output feedback identifier.