User Equipment Trust Reevaluation for Reliable ML Data Collection

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

The data collection process for machine learning model training in wireless communications is vulnerable to adversarial UEs that can inject perturbed measurements or updates, leading to compromised training performance and potential model misdirection.

Innovation Solution

Implement trust reevaluation mechanisms for UEs participating in data collection, triggered by specific events, to ensure only suitable data sets are used for training, thereby maintaining the integrity of machine learning models.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If data collection from all UEs is permitted for ML model training, then the quantity of training data increases, but the reliability of the training data decreases due to adversarial UEs injecting perturbed measurements

Engineering Contradiction:

Improvequantity of training dataVSAvoidreliability of training data

Core Design Contradiction:

Quantity of substanceVSReliability

Solution Approach 1:

The system performs preliminary trust evaluation of UEs before allowing them to participate in data collection for ML model training. This preliminary action identifies and flags potentially adversarial UEs before they can inject perturbed measurements, thus preserving data reliability while maintaining broad data collection from trusted sources

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

A network entity acts as an intermediary between UEs and the ML training process. This intermediary evaluates trustworthiness of UE data submissions and filters out potentially adversarial data before it reaches the training pipeline, enabling the system to maintain both high data quantity and high reliability through selective acceptance

Inventive Principle:

Principle #24Intermediary (Mediator)

2Reliability

If trust evaluation mechanisms are implemented for UEs, then the reliability of training data improves, but the device complexity increases due to additional evaluation infrastructure

Engineering Contradiction:

Improvereliability of training dataVSAvoidcomplexity of evaluation infrastructure

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The trust evaluation functionality is merged into existing network entities that are already part of the wireless communication infrastructure. By combining trust evaluation with existing network functions rather than introducing completely separate evaluation systems, the patent reduces overall device complexity while maintaining data reliability

Inventive Principle:

Principle #5Merging (Combining)

Solution Approach 2:

The network entity performing trust evaluation is designed to serve multiple functions: it evaluates UE trustworthiness, manages data collection permissions, and coordinates with ML training processes. This multi-functionality reduces the need for separate specialized components, thereby reducing overall system complexity while maintaining reliability

Inventive Principle:

Principle #6Universality (Multi-functionality)

Data Source

PatentUS12375940B2User equipment trust re-evaluation

Publication Date: 2025.07.29 QUALCOMM INC

AI summary

Certain aspects of the present disclosure provide a method for wireless communications at a first network entity, comprising: outputting, to a second network entity, a request for an evaluation of suitability of at least one user equipment (UE) for data collection for at least one of machine learning (ML) model training or scenario evaluation; and obtaining, in response to the request, results of the evaluation.