Search Query Modification Using Dynamic Personalized Knowledge Graphs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current search mechanisms fail to accurately understand user preferences, relationships, and intent based on past search history and current context, leading to irrelevant search results.
Innovation Solution
A method and system that modify search queries using a personalized knowledge graph dynamically generated from on-device data, including sensor data, settings, and application information, to enhance query understanding and provide relevant results.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If search mechanisms use only basic query processing without personalization, then system complexity is low, but search result relevance to user preferences and context is poor
Solution Approach 1:
The system performs preliminary actions by dynamically generating a personalized knowledge graph before processing the search query. This knowledge graph is built from user profile data, past search history, and contextual information, enabling the system to understand user preferences and intent in advance of the actual search operation
Solution Approach 2:
The personalized knowledge graph serves as an intermediary between the user's search query and the search results. It mediates by enriching the query with user-specific attributes and relationships, translating basic search terms into personalized search criteria that reflect user preferences, relationships, and context
2Adaptability or versatility
If search mechanisms use past search history and current context to understand queries, then search result personalization improves, but the system's ability to accurately understand user preferences and relationships deteriorates due to existing system limitations
Solution Approach 1:
The knowledge graph is dynamically generated and updated based on user interactions, profile changes, and new contextual information. The graph structure and relationships evolve over time, allowing the system to adapt to changing user preferences and accurately reflect current user state for improved preference understanding
Solution Approach 2:
The system adds another dimension to query processing by incorporating user-specific attributes, relationships, and contextual factors into the search process. This transforms the search from a simple keyword-matching operation into a multi-dimensional personalized search that considers user preferences, social relationships, and situational context
3Measurement precision
If the system modifies search queries based on user context and personalized knowledge graph, then search result relevance improves, but processing time and computational resources increase
Solution Approach 1:
The personalized knowledge graph is generated in advance and maintained for quick access during search operations. User profile data, relationships, and contextual information are pre-processed and structured in the knowledge graph, eliminating the need for complex real-time analysis during query processing
Solution Approach 2:
The system creates a simplified copy or representation of user preferences and context in the knowledge graph structure. This copied representation can be quickly queried and applied to modify search queries without requiring access to or processing of the entire user data set in real-time
Data Source
AI summary
A method and a modification system for modifying a search query for a user are provided. The method includes receiving at least one search query from a user; determining a present context of the user, based on the at least one search query and on-device data; identifying at least one relevant attribute related to the search query based on the present context of the user and a personalized knowledge graph, wherein the personalized knowledge graph is dynamically generated based on the on-device data; and modifying the search query, based on the at least one relevant attribute and a relation of the at least one relevant attribute with the search query, to obtain at least one modified search query.