access control in cyberspace information retrieval systems: an efficient method for security-aware semantic search

As the use of semantic search systems for information retrieval from massive data volumes grows, the need for effective access control has become a key challenge. traditional access control methods lack the necessary efficiency for vector-based architectures. this paper presents a simple and efficient method for implementing hierarchical access control in semantic search systems. our approach is based on appending a security dimension to text embedding vectors, embedding the access level of each document directly into its vector representation. we implemented and evaluated this method using the 20 newsgroups dataset, the all-mpnet-base-v2 language model, and the faiss library. experimental results show that this method correctly filters user access based on their permission levels, while its computational and storage overhead is nearly zero. this approach offers a practical and scalable solution for building intelligent and secure search systems.