Client-Side Persona Categorization for Privacy-Preserving Recommendations
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Information services face challenges in personalizing content for users whose data is not collected on servers due to privacy restrictions, leading to reduced accuracy and relevance of presented information, as well as increased data processing burdens on servers.
Innovation Solution
Generating persona and context categorization data offline using previously collected user data from consenting users, allowing client devices to determine persona traits and assign categories, thereby selecting relevant information without real-time server processing.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If user data is collected on servers for personalization, then accuracy of personalized recommendations is improved, but privacy restrictions and data processing burdens increase
Solution Approach 1:
The system performs preliminary action by generating persona categorization data offline using previously collected user data from consenting users. This pre-computed persona data is then made available to client devices for real-time use, eliminating the need for real-time server processing while maintaining personalization accuracy.
Solution Approach 2:
The invention extracts the essential personalization capability from real-time server processing and relocates it to client devices. By extracting persona categorization data and processing algorithms to the client side, the system eliminates privacy concerns associated with server-side data collection while preserving personalization functionality.
2Measurement precision
If real-time server processing is used for personalization, then relevance of presented information is improved, but server processing loads increase
Solution Approach 1:
The system performs preliminary action by pre-computing persona categorization data offline using previously collected user data. This pre-computed data is cached on client devices, enabling real-time personalization without requiring real-time server processing, thus reducing server processing loads while maintaining information relevance.
Solution Approach 2:
The invention creates a copy of the personalization processing capability on client devices. By copying persona categorization data and processing algorithms to the client side, the system enables local real-time processing that mirrors server functionality without actually loading the server, thereby reducing server processing loads.
3Productivity
If client devices process persona data locally, then server processing loads are reduced, but device computational resources are consumed
Solution Approach 1:
The system applies partial action by processing only the essential persona categorization data locally on client devices rather than all raw user data. The client devices perform limited computational tasks (categorization based on pre-computed personas) rather than full data processing, reducing device computational resource consumption while still achieving server load reduction.
Data Source
AI summary
A client device can identify user data pertaining to use of the client device by a user. The client device can determine at least one persona trait of the user based on the user data pertaining to the use of the client device by the user. The client device can receive persona categorization data, the persona categorization data specifying a plurality of persona categories and, for each persona category, a plurality of persona traits. Based on the at least one determined persona trait, the client device can assign the user to a persona category selected from the plurality of persona categories. Based on the persona category to which the user is assigned, the client device can identify information to present to users who are assigned to the persona category to which the user is assigned. The client device can present to the user the identified information.