Sublinear smart semantic search based on knowledge graph over encrypted database

Abstract

In light of the growing significance of data privacy in cloud computing, searchable encryption has emerged as a prominent area of research in recent years. Given that the majority of achievements are only concentrated on keyword-based search, there is a clear need to explore secure semantic extension to enhance the utilization of information. Nevertheless, majority of secure semantic schemes mainly focus on the direct combination of traditional searchable encryption and semantic extension techniques, which results in the high computation and communication overheads for multi-keyword queries. Besides, to enhance the flexibility and extensibility of secure semantic search, some researchers have already proposed encryption methods for structured data. However, the effectiveness and practicality of these methods above are often limited by their rigid structure description and redundant index building. To handle these problems, we present a secure sublinear semantic search scheme based on the knowledge graph, which can achieve sublinear knowledge graph (KG) search and smart sentence queries concurrently. The knowledge graph encryption index is constructed with the same structure as the inverted index, thereby enabling sublinear efficiency. Furthermore, the system also offers functions for both node and edge semantic extension in the knowledge graph. We demonstrate that our proposed scheme is adaptively secure and present a series of experiments to analyze its performance. The experimental results show that our scheme exhibits significant advantages in all computation, communication and expression situations.