Download PDFOpen PDF in browserLocal Privacy Protection System Framework Based on Encryption Algorithm LibraryEasyChair Preprint 66776 pages•Date: September 24, 2021AbstractHow to effectively protect the privacy of data, to establish a data center platform, solve the problem of deep learning vulnerable, has become a urgently need to solve the problem, in which the data is desensitization treatment is an effective way to avoid data privacy, this study puts forward a new solution, first identify sensitive data, then put the sensitive data from the original data. At the same time, the desensitization technology library of sensitive data is built, and then the sensitive data is desensitized. Finally, the data is transmitted to the model for training, so that the deep learning model is delivered for training without changing the original distribution of data, and the privacy leakage caused by the attack of the model is avoided. Keyphrases: Data Center Platform, Encryption Algorithm Library, deep learning, privacy protection
|