Supercharging Federated Intelligence Retrieval

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Stripelis, Dimitris, Foley, Patrick, Naseri, Mohammad, Lindskog-Münzing, William, Ng, Chong Shen, Beutel, Daniel Janes, Lane, Nicholas D.
Formato: Preprint
Publicado: 2026
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866918411016077312
author Stripelis, Dimitris
Foley, Patrick
Naseri, Mohammad
Lindskog-Münzing, William
Ng, Chong Shen
Beutel, Daniel Janes
Lane, Nicholas D.
author_facet Stripelis, Dimitris
Foley, Patrick
Naseri, Mohammad
Lindskog-Münzing, William
Ng, Chong Shen
Beutel, Daniel Janes
Lane, Nicholas D.
contents RAG typically assumes centralized access to documents, which breaks down when knowledge is distributed across private data silos. We propose a secure Federated RAG system built using Flower that performs local silo retrieval, while server-side aggregation and text generation run inside an attested, confidential compute environment, enabling confidential remote LLM inference even in the presence of honest-but-curious or compromised servers. We also propose a cascading inference approach that incorporates a non-confidential third-party model (e.g., Amazon Nova) as auxiliary context without weakening confidentiality.
format Preprint
id arxiv_https___arxiv_org_abs_2603_25374
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Supercharging Federated Intelligence Retrieval
Stripelis, Dimitris
Foley, Patrick
Naseri, Mohammad
Lindskog-Münzing, William
Ng, Chong Shen
Beutel, Daniel Janes
Lane, Nicholas D.
Information Retrieval
Computation and Language
Cryptography and Security
Machine Learning
68P20, 68T05, 62M45, 68P25, 68T50, 68T10
H.3.3; I.2.7
RAG typically assumes centralized access to documents, which breaks down when knowledge is distributed across private data silos. We propose a secure Federated RAG system built using Flower that performs local silo retrieval, while server-side aggregation and text generation run inside an attested, confidential compute environment, enabling confidential remote LLM inference even in the presence of honest-but-curious or compromised servers. We also propose a cascading inference approach that incorporates a non-confidential third-party model (e.g., Amazon Nova) as auxiliary context without weakening confidentiality.
title Supercharging Federated Intelligence Retrieval
topic Information Retrieval
Computation and Language
Cryptography and Security
Machine Learning
68P20, 68T05, 62M45, 68P25, 68T50, 68T10
H.3.3; I.2.7
url https://arxiv.org/abs/2603.25374