US 12,488,136 B1
Systems and methods for access control for federated retrieval-augmented generation
Shaunak Godbole, San Francisco, CA (US); Shih Ping Chang, San Francisco, CA (US); Clemens Mewald, Lafayette, CA (US); and Anant Bhardwaj, San Francisco, CA (US)
Assigned to Instabase, Inc., Dover, DE (US)
Filed by Instabase, Inc., Dover, DE (US)
Filed on Mar. 29, 2024, as Appl. No. 18/622,047.
Int. Cl. G06F 21/00 (2013.01); G06F 21/62 (2013.01)
CPC G06F 21/6227 (2013.01) 20 Claims
OG exemplary drawing
 
1. A system configured to provide access control for federated retrieval-augmented generation (FRAG) using a set of vector databases including a first vector database and a second vector database, wherein the first vector database stores a first set of vector embeddings representing content within a first set of electronic files, and a second set of vector embeddings representing content within a second set of electronic files, the system comprising:
one or more hardware processors configured by machine readable instructions to:
receive, from a user identified by a user identifier, user input representing a user query, wherein the user query entails a set of tasks, wherein the set of tasks includes at least a first task associated with the first set of electronic files and a second task associated with the second set of electronic files;
verify whether the user has access to the first set of electronic files and the second set of electronic files, wherein verification is based on the user identifier;
responsive to the verification of the user having access to the first set of electronic files and the second set of electronic files, generate a first agent query based on the first task and a second agent query based on the second task;
responsive to verification of the user having access to the first set of electronic files, select a first subset of the first set of vector embeddings based on the first agent query;
responsive to the verification of the user having access to the first set of electronic files, provide the first subset of the first set of vector embeddings to one or more machine learning models as context for a first prompt based on the first agent query;
responsive to the verification of the user having access to the first set of electronic files, provide the first prompt to the one or more machine learning models;
obtain a first reply from the one or more machine learning models in response to the first prompt;
responsive to verification of the user having access to the second set of electronic files, select a second subset of the second set of vector embeddings based on the second agent query;
responsive to the verification of the user having access to the second set of electronic files, provide the second subset of the second set of vector embeddings to the one or more machine learning models as context for a second prompt based on the second agent query;
responsive to the verification of the user having access to the second set of electronic files, provide the second prompt to the one or more machine learning models;
obtain a second reply from the one or more machine learning models in response to the second prompt;
generate a reply to the user query, wherein the reply is based on the first reply and the second reply; and
effectuate a presentation of the reply to the user, through a user interface.