Privacy Protected Dynamic User Clustering Middleware

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing methods for privacy preservation in personalized services and data analytics often compromise user privacy due to resource-intensive homomorphic encryption and centralized solutions that can be vulnerable to malicious actors, especially in dynamic environments where user clusters change frequently.

Innovation Solution

Implementing a middleware node system that dynamically clusters users based on common interests, splitting clusters into sub-clusters when the user count exceeds a predetermined threshold to ensure a minimum cluster size, thereby maintaining user privacy and improving recommendation relevance.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If homomorphic encryption is used for privacy preservation in data analytics, then user privacy is protected, but computational resources and processing time are excessively consumed

Engineering Contradiction:
Improveprivacy protectionVSAvoidcomputational resources
Core Design Contradiction:
ReliabilityVSUse of energy by moving object

Solution Approach 1:

The patent extracts the privacy protection mechanism from centralized homomorphic encryption and implements it at the distributed edge devices. Each edge device performs local clustering and filtering operations, removing the need for resource-intensive centralized encryption processing while maintaining privacy guarantees through distributed computation architecture.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent introduces trusted execution environments (TEEs) and secure multi-party computation protocols as intermediary layers between data sources and analytics systems. These intermediaries enable privacy-preserving data sharing and collaborative filtering without requiring full homomorphic encryption, reducing computational overhead while maintaining security guarantees.

Inventive Principle:
Principle #24Intermediary (Mediator)

2Device complexity

If centralized solutions are used for data analytics, then implementation is simplified, but vulnerability to malicious actors increases

Engineering Contradiction:
Improvesystem implementationVSAvoidvulnerability to malicious actors
Core Design Contradiction:
Device complexityVSObject-affected harmful factors

Solution Approach 1:

The patent segments the centralized analytics system into distributed edge computing nodes that perform local clustering and data filtering. This segmentation distributes trust across multiple independent entities, preventing any single malicious actor from compromising the entire system while maintaining implementation simplicity through modular architecture.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent implements preliminary privacy-preserving operations at the edge devices before data is transmitted to centralized analytics platforms. Clustering, filtering, and aggregation are performed in advance, so that even if the centralized system is compromised, the sensitive raw data has already been transformed into anonymized insights that cannot be traced back to individual users.

Inventive Principle:
Principle #10Preliminary action

3Measurement precision

If user clusters are kept large for data analytics, then statistical accuracy improves, but user identification by malicious parties becomes easier

Engineering Contradiction:
Improvestatistical accuracyVSAvoiduser identification risk
Core Design Contradiction:
Measurement precisionVSObject-affected harmful factors

Solution Approach 1:

The patent implements dynamic cluster formation where user groups are continuously reconfigured based on real-time behavior patterns and contextual factors. Clusters are made dynamic through frequent updates and transformations, making it difficult for malicious actors to track or identify specific users over time while maintaining sufficient cluster sizes for statistically accurate analytics.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent applies different privacy protection strategies to different clusters based on their size, sensitivity, and risk profile. High-risk clusters receive enhanced protection through differential privacy mechanisms and noise addition, while lower-risk clusters use lighter protection methods, optimizing the balance between statistical accuracy and privacy protection locally for each cluster.

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP2725538B1Privacy protected dynamic clustering of end users
Publication Date: 2017.12.13 ALCATEL LUCENT SA
  • EP2725538B1 patent drawingFigure 1(a)
  • EP2725538B1 patent drawingFigure 1(b)
  • EP2725538B1 patent drawingFigure 2

AI summary

The present subject matter discloses a method for privacy protected dynamic clustering of end users. In one implementation, a number of end users subscribed to a cluster are ascertained, with the cluster being indicative of a group of end users having a common interest. The method further includes splitting the cluster into two or more sub-clusters if the ascertained number of end users is greater than a predetermined explore-splitting threshold and number of end users subscribed to each of the two or more sub-clusters is greater than a predetermined cluster size threshold.