应用于人工智能生成内容的文本处理方法、装置及介质
By identifying and replacing the information to be de-identified in the text to be processed on the user terminal, generating de-identified text and sending it to the cloud server for processing, and restoring the information after receiving it, the problem of information leakage caused by large cloud NLP models is solved, and information protection and restoration are achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- TENCENT CLOUD COMPUTING (BEIJING) CO LTD
- Filing Date
- 2023-02-23
- Publication Date
- 2026-07-17
AI Technical Summary
When users use large NLP models deployed on cloud servers to perform NLP tasks on text to be processed, it can easily lead to the leakage of important user information.
By identifying the information to be de-identified in the text to be processed, a de-identification mapping table is generated. The information to be de-identified in the text to be processed is replaced with the de-identified information to obtain the de-identified text, which is then sent to the AI-generated content server for text processing. After receiving the output text, the information to be restored is identified and replaced according to the de-identified information to obtain the target text.
It protects users' important information from being leaked and enables the restoration of information from the output text, thus improving the security of de-identified text.
Smart Images
Figure CN116186771B_ABST