Application Function QoE Provisioning Through Vertical Federated Learning

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing cellular network technologies face challenges in collecting real-time quality of experience (QoE) data across different domains due to privacy and commercial concerns, leading to inefficient network resource provisioning and increased signaling overhead.

Innovation Solution

Implementing vertical federated learning (VFL) techniques to enable collaborative training of machine learning models across user equipment (UE), application function (AF), and network data analytics function (NWDAF) without sharing raw data, using localized models to generate intermediate results and optimize QoE metrics.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If real-time QoE data is collected across different domains, then QoE prediction accuracy is improved, but privacy risks and signaling overhead increase

Engineering Contradiction:

ImproveQoE prediction accuracyVSAvoidprivacy risks and signaling overhead

Core Design Contradiction:

Measurement precisionVSObject-affected harmful factors

Solution Approach 1:

The patent segments the data collection and processing system into multiple independent domains (UE domain with AF, network domain with NWDAF, and model domain). Each domain processes data locally and only shares aggregated results or model parameters, not raw data. This segmentation enables accurate QoE prediction through collaborative processing while minimizing privacy risks and signaling overhead by avoiding centralized raw data collection.

Inventive Principle:

Principle #1Segmentation

2Productivity

If collaborative training of machine learning models is implemented across multiple domains, then QoE metric optimization is improved, but data sharing complexity increases

Engineering Contradiction:

ImproveQoE metric optimizationVSAvoiddata sharing complexity

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The patent introduces an intermediary mechanism where model parameters and aggregated statistics serve as mediators between domains. Instead of directly sharing complex raw data, each domain processes data locally and exchanges simplified model parameters through the NWDAF. This intermediary approach enables collaborative training and QoE optimization while significantly reducing data sharing complexity and protecting domain-specific data privacy.

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentUS20250310215A1Application function influenced network and quality of experience provisioning

Publication Date: 2025.10.02 APPLE INC

AI summary

Techniques are described herein for paging adaption. An example, method includes receiving, from a user equipment (UE), a request for configuration optimization information for an application. The method can further include causing a first machine learning model at an application function (AF) to generate first intermediate results based at least in part on the request and collaborative analytics information from the UE. The method can further include accessing second intermediate results generated using a second machine learning model at a network data and analytics function (NWDAF). The method can further include causing a third machine learning model at the AF to predict a quality of experience (QoE) label metric associated with the application based at least in part on the first intermediate results and the second intermediate result.