Sharable Persona Models for Privacy-Controlled Online Personalization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing online services and applications individually build and control user profiles, leading to inefficiencies in user experience customization, privacy concerns, and increased training time for new applications, as users lack control over their profile data and privacy settings.
Innovation Solution
A personalization model that allows users to create and manage sharable personas across different ecosystems, enabling control over data exposure, context-aware content delivery, and optimized network usage, while ensuring privacy and security through user-owned personas.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If online services individually build and control user profiles, then each service can customize user experience, but training time increases and user privacy control decreases
Solution Approach 1:
The patent implements pre-trained persona templates that contain pre-existing knowledge and behavior patterns. When a user joins a new service, these pre-trained personas can be immediately deployed or quickly adapted, eliminating the need for extensive re-training from scratch. The persona templates are prepared in advance with general user characteristics that can be applied across multiple services.
Solution Approach 2:
The patent creates user personas that are universal and transferable across different online services and ecosystems. A single persona instance can serve multiple services simultaneously, allowing the user profile to be reused without requiring separate training for each service. This multi-functional persona approach reduces redundant training while maintaining service-specific customization capabilities.
2Ease of operation
If online services individually build and control user profiles, then each service has full control over profile data, but user privacy control and data security decrease
Solution Approach 1:
The patent introduces user personas as an intermediary layer between users and online services. Personas act as controlled avatars that users can manage, update, and revoke at will. This intermediary structure gives users direct control over what information is shared with services, while services interact with the persona rather than directly accessing sensitive user data. Users can grant or revoke persona access to specific services without exposing their actual identity or personal information.
Solution Approach 2:
The patent segments user identity and profile data into distinct persona instances that can be independently managed. Instead of a single monolithic user profile, users can create multiple personas with different characteristics and data exposure levels for different services. This segmentation allows users to control privacy on a service-by-service basis, limiting data exposure to only what is necessary for each specific service while maintaining profile management capabilities.
3Measurement precision
If user profiles are built based on observed interactions, then personalization accuracy improves, but network bandwidth utilization and power consumption increase
Solution Approach 1:
The patent pre-trains persona templates with general user behavior patterns and preferences before deployment. This preliminary training phase occurs offline or during initial setup, allowing the persona to have baseline personalization capabilities without requiring continuous online learning. The pre-trained persona can provide accurate personalization from the start, reducing the need for frequent data synchronization and re-training operations that consume network bandwidth and device power.
4Adaptability or versatility
If each online service builds individual user profiles, then service-specific personalization is achieved, but network bandwidth utilization increases due to duplicative training
Solution Approach 1:
The patent implements universal persona templates that can be deployed across multiple online services simultaneously. Instead of each service training separate profiles, a single persona instance serves multiple services, eliminating duplicative training operations. The persona maintains service-specific personalization capabilities while being trained once, significantly reducing network bandwidth consumption for data synchronization and model updates across the ecosystem.
Data Source
AI summary
A cloud-based personalization service utilizes a personalization model providing service users with tools to create, personalize, and manage their own user profiles, which are abstracted in the form of “personas,” that can learn from, and be shared across, different ecosystems of online services and applications. One or more personas may be locally instantiated on the user's various computing devices by a personalization system. The personalization service and system interoperate to enable users to set operating parameters, preferences, and content-filtering criteria for their personas. Users can strictly control the information and parameters that their personas expose to the services and applications to protect user privacy and enhance the quality of online interactions. The personalization model facilitates personalization of the user's personas to individual online services and applications based on context-awareness to further improve the relevance of delivered content and user experiences.