US 12,423,332 B1
Enhanced distributed retrieval augmented generation system
William Jeffery White, Plano, TX (US); Xin Ma, Pflugerville, TX (US); and George Currie, Canyon Lake, TX (US)
Assigned to Dell Products L.P., Round Rock, TX (US)
Filed by Dell Products L.P., Round Rock, TX (US)
Filed on Nov. 29, 2024, as Appl. No. 18/963,965.
Claims priority of provisional application 63/691,160, filed on Sep. 5, 2024.
Int. Cl. G06F 16/00 (2019.01); G06F 16/28 (2019.01)
CPC G06F 16/285 (2019.01) 20 Claims
OG exemplary drawing
 
1. A method of forming query optimized representations (QORs), comprising:
normalizing enterprise data received at an enhanced distributed retrieval augmented generation (RAG) system, wherein:
the enhanced distributed RAG system is located on a distributed edge site of a plurality of distributed edge sites,
each distributed edge site of the plurality of distributed edge sites has a different enhanced distributed RAG system of a plurality of enhanced distributed RAG systems, and
the plurality of distributed edge sites are connected to a centralized system and the centralized system comprises a global federated RAG module;
extracting relevant information from the enterprise data;
chunking the relevant information into relevant information parts;
performing semantic tagging of the relevant information parts;
creating a local QORs of the relevant information parts;
transmitting the relevant information, the relevant information parts, semantic tagging data, and the local QORs to a local data cache of a distributed data management system connected to the enhanced distributed RAG system;
transmitting the local QORs to the global federated RAG module; and
transmitting the local QORs to a local small foundation system in response to a first query received at the local small foundation system, wherein the local QORs is used to form prompts based on the first query for a generative artificial intelligence model.