Multi-party joint vector knowledge base retrieval method and system for privacy protection

The invention provides a privacy-protected multi-party joint vector knowledge base retrieval method and system, and the method comprises the steps: multi-party text corpora construct indexes and respectively upload the indexes to a trusted third party, the trusted third party distributes random protection secret parameters, and each party generates auxiliary data in combination with local random parameters; each party obtains an embedded vector of the corpus, and data subjected to dimension reduction processing and random homogeneous transformation and auxiliary data are uploaded to a joint vector database server side; a user carries out corpus embedding on a query request text and then carries out dimensionality reduction and random homogeneous transformation to generate a vector to be retrieved; performing similarity retrieval after the vector to be retrieved and auxiliary data are operated in the joint vector database, and returning an index result; and after the vector similarity retrieval is completed, q