.png){kind=icon}
.png){kind=icon}
Return
Multiuser Personalized Ciphertext Retrieval Scheme Based on Deep Learning
.png){kind=icon}
PRE
AI
DOI：10.1109/JIOT.2023.3305359.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
With the rapid development of cloud computing technology and Internet of Things (IoT), enterprises and organizations tend to outsource local data to cloud servers and use searchable encryption (SE) technology to access and search encrypted data. However, the existing symmetric SE (SSE) schemes pay less attention to multiuser environments and users' interest, which cause poor experience to users. In this article, we propose a multiuser personalized ciphertext search scheme (MPCS) by extending deep learning technology and SSE technology, which can achieve personalized retrieval and multiuser retrieval at the same time. MPCS achieves secure transmission of document keys and fine-grained access control by combining matrix decomposition and ciphertext policy attribute-based encryption (CP-ABE) technology, which assigns different private keys to each authorized user in the system and allows setting different access rights for different users. Second, we build an interest model for different users and design a user query update algorithm based on the attention mechanism to provide personalized ranking results, which improves the retrieval experience of users. In addition, the computation overhead of MPCS is lightweight, the size of ciphertext and key will not increase linearly with the number of attribute values, and MPCS supports lightweight document updates, which greatly reduces the computation overhead of the system. Formal security analysis verifies the security of MPCS, and simulation experiments on real data sets show that MPCS is feasible and efficient in practice.
Keywords:
Deep learning
Access control
multiuser
personalize
privacy protection
searchable encryption (SE)
Journal
IF:
8.9
Papers:
1.4W
Citations:
7.8W