Electronic Apparatus for Private-Attribute Masking in Speech Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Natural language processing models that utilize personal information face a risk of exposing sensitive user data, leading to potential leaks during transmission or storage.
Innovation Solution
An electronic apparatus and method that identifies and separates personal data into private and public attributes, transmitting only public attributes to a server while masking or omitting private attributes, thereby enhancing security and maintaining data accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If personal information is transmitted to the server for natural language processing, then processing accuracy is improved, but data security deteriorates
Solution Approach 1:
The personal information is segmented into first sub data (private attributes) and second sub data (public attributes). This segmentation allows the system to transmit only the necessary public attributes to the server for processing, while keeping private attributes locally, thus resolving the contradiction between processing accuracy and data security.
Solution Approach 2:
The invention extracts and transmits only the essential second sub data (public attributes) needed for natural language processing to the server, while retaining the first sub data (private attributes) locally. This extraction approach maintains processing accuracy by providing necessary context while eliminating data exposure risks associated with transmitting sensitive information.
2Productivity
If personal information is stored in external device, then processing capability is improved, but reliability deteriorates
Solution Approach 1:
By segmenting personal information into private and public attributes, the system can store only public attributes in the external device (server), reducing the reliability risk while maintaining processing capability. The private attributes remain in the terminal device where they can be securely managed.
Solution Approach 2:
The invention introduces an intermediary mechanism that processes natural language using only public attributes on the server, without requiring direct access to sensitive private information. This intermediary approach maintains processing capability while improving reliability by eliminating the need to store sensitive data externally.
Data Source
AI summary
An electronic apparatus includes a memory configured to store personal data associated with a user, a communication interface configured to communicate with a server, and a processor configured to obtain, from the personal data associated with the user stored in the memory and based on a user speech being received, personal data corresponding to the user speech, identify, from the personal data corresponding to the user speech, first sub data corresponding to a private attribute and second sub data corresponding to a public attribute, obtain combination information based on the user speech and the second sub data, transmit, through the communication interface, the combination information to the server, and obtain response information corresponding to the combination information from the server.