Publication: Order-preserving Encryption as a Tool for Privacy-Preserving Machine Learning
Дата
2020
Авторы
Journal Title
Journal ISSN
Volume Title
Издатель
Аннотация
© 2020 IEEE.An order-preserving encryption is an encryption scheme based on strictly increasing functions. It allows mapping a set of plaintext into a set of ciphertexts with the order relation specified on a set of plaintexts. Thus, data comparisons and search for the minimum or maximum elements become possible without decrypting data. This type of encryption is mainly used to protect cloud databases in the case when it is necessary to make queries to them. However, it is almost not used in privacy-preserving data mining and machine learning. Nevertheless, it is possible to use this type of encryption in privacy-preserving machine learning, but only for certain algorithms. In this paper, we consider some existing order-preserving encryption schemes and suggest some cases of machine learning, where they can be applied to obtain correctly working privacy-preserving machine learning algorithms. An explanation is also given how and why it is possible to apply order-preserving encryption to these algorithms.
Описание
Ключевые слова
Цитирование
Lisin, N. Order-preserving Encryption as a Tool for Privacy-Preserving Machine Learning / Lisin, N., Zapechnikov, S. // Proceedings of the 2020 IEEE Conference of Russian Young Researchers in Electrical and Electronic Engineering, EIConRus 2020. - 2020. - P. 2090-2092. - 10.1109/EIConRus49466.2020.9039294