Local Knowledge Graph Update for Privacy and Long-Term Behavior
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for generating knowledge graphs face challenges in storing vast amounts of data and transmitting sensitive personal information, leading to incomplete knowledge graphs and inability to utilize long-term user behavior patterns, as well as generating personalized knowledge graphs.
Innovation Solution
An electronic apparatus that selectively identifies user data, acquires context information, and updates a knowledge graph using personalized information stored locally, applying weights and new entity information to incorporate relevant data patterns and behaviors, thereby generating an optimized knowledge graph without external server transmission of sensitive information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional methods transmit knowledge data to external servers for generating knowledge graphs, then knowledge graph generation capability is improved, but data security and privacy protection deteriorate due to transmission of sensitive personal information
Solution Approach 1:
The patent introduces an intermediary processing mechanism where the knowledge graph generation process is mediated through local processing of personal information. Instead of directly transmitting sensitive data to external servers, the system uses intermediate processing steps that extract necessary knowledge patterns while preserving privacy, thus resolving the contradiction between generation capability and data security
Solution Approach 2:
The patent extracts only the essential knowledge patterns and relationships from user data without transmitting the raw personal information itself. By separating the extraction of knowledge structures from the transmission of sensitive data, the system achieves knowledge graph generation while maintaining data security
2Quantity of substance
If conventional methods delete past data to manage storage limitations, then storage capacity is improved, but long-term user behavior pattern analysis deteriorates
Solution Approach 1:
The patent segments data storage into multiple hierarchical levels: raw data storage, processed knowledge data storage, and pattern template storage. This segmentation allows the system to retain essential historical information in compressed form while managing storage resources efficiently, preventing loss of long-term behavior patterns
Solution Approach 2:
The patent transforms raw user data into processed knowledge data with different parameters - converting detailed personal information into abstracted knowledge patterns and relationships. This parameter transformation reduces storage requirements while preserving the essential information needed for long-term pattern analysis
3Measurement precision
If the electronic apparatus processes all user data to generate personalized knowledge graphs, then personalization accuracy is improved, but processing complexity and resource consumption worsen
Solution Approach 1:
The patent applies partial processing by focusing computational resources on processing only the user data that is most relevant to knowledge graph generation. Instead of uniformly processing all data, the system selectively processes data that contributes most to personalization accuracy, reducing overall processing complexity
Solution Approach 2:
The patent performs preliminary processing of user data into structured knowledge data before generating the knowledge graph. By pre-processing and organizing data into relevant patterns and relationships in advance, the system reduces the complexity of the final graph generation process while maintaining personalization accuracy
Data Source
AI summary
An electronic apparatus is disclosed. The electronic apparatus includes: a communication interface comprising communication circuitry, a memory including a knowledge graph including a plurality of knowledge data and at least one command, and a processor connected with the memory and configured to control the electronic apparatus, wherein the processor is configured, by executing the at least one command, to: receive user data from at least one external apparatus through the communication interface, identify whether a user behavior occurred based on the user data, based on identifying the user behavior, acquire context information related to the user behavior, identify first knowledge data indicating relevance among information on the user behavior, the context information, and personalized information stored in the memory, and compare the first knowledge data and second knowledge data included in the knowledge graph included in the memory and update the knowledge graph, wherein the first knowledge data and the second knowledge data include at least one entity information of the knowledge graph and information regarding a relation between the at least one entity information.