analiza aiqbench
This paper proposes applying one shared random projection to both datastore and query embeddings in RAG, aiming to reduce leakage of sensitive retrieval data. Its main trade-off is that the privacy gain is empirical, not formal: the claimed provable guarantee covers similarity preservation, while the privacy comparison to differential privacy methods depends on specific attack models. Readers building privacy-sensitive RAG systems should treat it as a strong, cheap baseline to validate against their own threat models rather than a substitute for formal guarantees.
Private Retrieval Augmented Generation via Random Projection
arXiv:2610.22297v1 Announce Type: new Abstract: Retrieval-Augmented Generation (RAG) enhances the capabilities of large language models (LLMs) by querying external structured knowledge. However, it can also introduce privacy risks by leaking sensitive information from the retrieval datastore. We propose a simple method to preserve datastore privacy in RAG systems via random projection. By applying the same projection to both datastore embeddings and query embeddings, our method provably preserves semantic similarity between queries and retrieved items while substantially mitigating data extraction attacks. Across multiple RAG architectures and datasets, we show that this lightweight approach achieves superior retrieval and generation performance compared to prior methods which leverage formal differential privacy (DP) guarantees, while exhibiting comparable empirical privacy under strong attack models. Our results for the first time suggest that random projection can serve as a competitive and practical baseline for privacy-preserving RAG systems.
Abstract:Retrieval-Augmented Generation (RAG) enhances the capabilities of large language models (LLMs) by querying external structured knowledge. However, it can also introduce privacy risks by leaking sensitive information from the retrieval datastore. We propose a simple method to preserve datastore privacy in RAG systems via random projection. By applying the same projection to both datastore embeddings and query embeddings, our method provably preserves semantic similarity between queries and retrieved items while substantially mitigating data extraction attacks. Across multiple RAG architectures and datasets, we show that this lightweight approach achieves superior retrieval and generation performance compared to prior methods which leverage formal differential privacy (DP) guarantees, while exhibiting comparable empirical privacy under strong attack models. Our results for the first time suggest that random projection can serve as a competitive and practical baseline for privacy-preserving RAG systems.
Artificial Intelligence (cs.AI); Cryptography and Security (cs.CR)
arXiv:2610.22297 [cs.AI]
(or arXiv:2610.22297v1 [cs.AI] for this version)
https://doi.org/10.48550/arXiv.2610.22297 Focus to learn more arXiv-issued DOI via DataCite (pending registration)
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Text integral sub licența deschisă a publicației · 233 cuvinte
Subiecte conexe
Spruce: Scalable Private Outsourced Retrieval Using Compact Embeddings
arXiv · 4 sept. 2026
Pointing the Way, Hiding the Destination: Practical Private Dense Retrieval at Scale
arXiv · 27 aug. 2026
TurboVec: A Case Study in Cost-Efficient Private Retrieval for Enterprise RAG via Codebook-Oblivious Quantization
arXiv · 21 iul. 2026
Certified Adaptive Refresh: Anytime-Valid Monitoring for Federated Conformal RAG
arXiv · 30 sept. 2026