Probabilistic Identity Database for Cross-Device User Identification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
There is no effective solution to identify and analyze user relationships across multiple Internet-connected devices, as users often utilize different devices with various identifiers and access points, making it challenging for companies to understand and reach individuals and families consistently.
Innovation Solution
The implementation of an identity database (IDB) system that uses a hierarchy of object identifiers, including household, user, and physical device identifiers, to create a graph structure and relationship tables, allowing for probabilistic inference of user relationships and device associations through clustering and scoring processes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If users communicate anonymously over the Internet without disclosing personal identifiers, then user privacy is protected, but the ability to identify users across different devices is lost
Solution Approach 1:
The patent introduces probabilistic identity objects as intermediary entities that link device identifiers to user identities without directly exposing personal identifiers. These probabilistic identity objects serve as mediators that enable user identification across devices while maintaining privacy by not requiring direct disclosure of personal information.
Solution Approach 2:
The system transforms deterministic identifier matching into probabilistic identity resolution by changing the parameter from exact match to probability-based association. This allows the system to work with anonymous and pseudonymous identifiers while still achieving reliable user identification through statistical methods.
2Adaptability or versatility
If different device identifiers and network addresses are used for each device, then device diversity and user flexibility are improved, but the complexity of identifying user relationships across devices increases
Solution Approach 1:
The patent segments the user identification system into hierarchical levels: device identifiers, probabilistic identity objects, and user identities. This segmentation allows each layer to handle specific aspects of identification independently, reducing overall system complexity while supporting diverse device usage.
Solution Approach 2:
The probabilistic identity objects serve as universal intermediaries that can associate any device identifier with user identities. This multi-functional component handles various identification scenarios across different devices, networks, and communication contexts without requiring device-specific identification mechanisms.
3Reliability
If multiple email accounts and pseudonyms are used for different devices, then user privacy and device separation are maintained, but the ability to analyze user-user and user-family relationships is degraded
Solution Approach 1:
Probabilistic identity objects act as intermediaries that link multiple email accounts and pseudonyms to underlying user identities and family relationships. These intermediaries enable relationship analysis by connecting anonymous identifiers to real identities through probabilistic associations without requiring direct identity disclosure.
Solution Approach 2:
The system uses feedback from observed device usage patterns, network connections, and interaction data to continuously refine probabilistic identity associations. This feedback mechanism improves relationship analysis accuracy over time by learning from actual user behavior while maintaining privacy protection.
Data Source
AI summary
A storage device including an information structure produced by a method comprising: determining unique cluster names during a time interval; creating, in a storage device, one or more relationship tables, wherein for each unique cluster name, creating one or more relationship tables includes, creating one or more pairings that each includes an individual object identifier member of a cluster corresponding to the unique cluster name and includes a produced association score for the individual object identifier member, and wherein for each unique cluster name, creating one or more relationship tables includes creating a relationship between a unique key name that matches the unique cluster name and each of the one or more pairings created for the unique cluster name; and repeating the acts of determining and updating at selectable time intervals.