Personalization Content Management with Offline-Situational Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing content delivery services face processing and network latencies due to dynamic determination of personalization information at runtime, limiting their ability to provide individualized personalization and often rely on limited offline approaches that neglect contextual information.
Innovation Solution
A system that precomputes offline personalization information and dynamically determines situational personalization based on context information using personalization models, such as weighted decision trees and neural networks, to generate personalized content efficiently.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If personalization information is dynamically determined at runtime, then personalization is adapted to user context, but processing latency and network latency increase
Solution Approach 1:
The patent precomputes and stores personalization information in offline personalization data structures before runtime needs arise. User profiles, preferences, and personalization rules are prepared in advance and cached, allowing the system to retrieve precomputed personalization data during runtime instead of computing it dynamically, thus reducing processing latency while maintaining personalization adaptability
Solution Approach 2:
The patent divides personalization information into two segments: offline personalization data (precomputed user profiles, preferences, and rules) and situational personalization data (context-specific information retrieved at runtime). This segmentation allows the system to handle different types of personalization requirements efficiently, using precomputed data for stable user attributes and dynamic retrieval for context-specific personalization
2Adaptability or versatility
If extensive personalization is provided, then user experience is improved, but system complexity and processing requirements increase
Solution Approach 1:
Complex personalization computations, including user profile analysis, preference extraction, and personalization rule application, are performed in advance during offline processing. The results are stored in structured data formats that simplify runtime retrieval and application, reducing the computational burden and system complexity during runtime while enabling extensive personalization
Solution Approach 2:
The patent creates simplified copies of user profiles and personalization data in optimized data structures that are easier to process at runtime. Instead of working with raw, complex user data, the system uses pre-processed copies containing extracted features and computed attributes, reducing processing complexity while maintaining personalization quality
3Loss of time
If offline personalization approaches are used, then processing latency is reduced, but contextual information is neglected
Solution Approach 1:
The patent segments personalization data into offline personalization information (precomputed user profiles and preferences) and situational personalization information (context-specific data retrieved at runtime). This segmentation enables the system to combine the speed benefits of offline processing with the relevance benefits of contextual information by retrieving situation-specific data only when needed
Solution Approach 2:
The patent merges offline personalization data with situational personalization data during runtime to create comprehensive personalized content. The system combines precomputed user profiles with context-specific information such as current location, device type, and situational preferences, achieving both fast processing and contextual relevance
Data Source
AI summary
The disclosure herein pertains to a system and method for management of personalization content. The system and method divide the personalization information into offline personalization information and situational personalization information. Offline personalization information is independent of context and predetermined before a content request. A personalization model can dynamically allocate the selection between offline personalization information and situational personalization information.