No abstract

SummaryAccess control can effectively protect users' data and privacy in social networks. Different access control models address different data privacy requirements, and the corresponding implementation technologies and performance are also quite different. The existing access control mechanisms of social networks rarely take into account the individual preferences of users, and cannot provide personalized services for users. Therefore, we have studied an analysis mechanism of user preference to quantify user preferences, and then use user preference values and some key information of users' social resources as decision vectors, and proposed an access control mechanism based on user preference (ACMBUP). ACMBUP turns the problem of access decision‐making into a problem of classification, that is, whether to allow access or not to allow access. At the same time, we add an encryption and decryption mechanism in ACMBUP to prevent data leakage and protect data when users interact with access mechanisms. Experiments show that this mechanism can automatically generate appropriate access control policies to meet the potential privacy needs of different users, so as to better protect the privacy of social networks data.

No abstract

With the advent of the multimedia era, the identification of sensitive information in social data of online social network users has become critical for maintaining the security of network community information. Currently, traditional sensitive information identification techniques in online social networks cannot acquire the full semantic knowledge of multimodal data and cannot learn cross-information between data modalities. Therefore, it is urgent to study a new multimodal deep learning model that considers semantic relationships. This paper presents an improved multimodal dual-channel reasoning mechanism (MDR), which deeply mines semantic information and implicit association relationships between modalities based on the consideration of multimodal data fusion. In addition, we propose a multimodal adaptive spatial attention mechanism (MAA) to improve the accuracy and flexibility of the decoder. We manually annotated real social data of 50 users to train and test our model. The experimental results show that the proposed method significantly outperforms simple multimodal fusion deep learning models in terms of sensitive information prediction accuracy and adaptability and verifies the feasibility and effectiveness of a multimodal deep model considering semantic strategies in social network sensitive information identification.