Local privacy protection system framework based on encryption algorithm library
Yingcheng Gu, Yongqiu Chen, Mingsheng Xu · International Conference on Mechanisms and Robotics (ICMAR 2022) · 2022
How to effectively protect the privacy of data, establish a data center platform and solve the problem of vulnerable deep learning has become an urgent problem to be solved. In particular, desensitization of data is an effective way to avoid data privacy disclosure. This study puts forward a new solution. Firstly, this article identifies the sensitive data, then extracts the sensitive data from the original data, constructs the desensitization technology library of the sensitive data, then desensitizes the sensitive data, and finally transfers the data to the model for training, so as to deliver the deep learning model for training without changing the original distribution of the data, it avoids the problem of privacy disclosure caused by model attack.