EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Private-RAG: A Privacy-Preserving Retrieval-Augmented Generation Method for Large Model Inference

Qianren Yang, Yong Li, Xiang Ma

Electronics · 10.06.2026 · 15 (12) · 2567

DOI: 10.3390/electronics15122567 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Electronics, online erschienen am 10.06.2026. Verfasst von Qianren Yang und 2 weiteren Autor:innen. Die Arbeit nennt 34 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Xinjiang Tianshan Talent Training Program.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Retrieval-augmented generation improves the factual consistency, knowledge timeliness, and scenario adaptability of large model inference services by incorporating external knowledge. However, it also introduces structural privacy risks, including private-knowledge leakage, prompt injection, and progressive information extraction in multi-turn interactions. To address these issues, this paper proposes Private-RAG, a privacy-preserving retrieval-augmented generation method for large model inference. The method constructs a composite threat model and a quantitative evaluation framework for the RAG pipeline, and further develops a layered collaborative defense mechanism consisting of controlled retrieval, sensitivity-aware context minimization, structured prompt isolation, and multi-criterion output gating. In addition, a risk feedback-driven budget accounting method is introduced to enable dynamic risk control in multi-turn interaction scenarios. Experimental results show that Private-RAG effectively reduces private-knowledge leakage, improves robustness against prompt injection, and suppresses cumulative privacy exposure while maintaining question-answering utility and a controllable deployment latency (e.g., 1165 ms), demonstrating superior privacy protection and inference robustness.

Förderung

  • Xinjiang Tianshan Talent Training Program

Zitieren

APA

Yang, Q., Li, Y., & Ma, X. (2026). Private-RAG: A Privacy-Preserving Retrieval-Augmented Generation Method for Large Model Inference. Electronics, 15(12), 2567. https://doi.org/10.3390/electronics15122567

Vancouver

Yang Q, Li Y, Ma X. Private-RAG: A Privacy-Preserving Retrieval-Augmented Generation Method for Large Model Inference. Electronics. 2026;15(12):2567. doi:10.3390/electronics15122567

BibTeX
@article{Yang2026,
  author = {Yang, Qianren and Li, Yong and Ma, Xiang},
  title = {Private-RAG: A Privacy-Preserving Retrieval-Augmented Generation Method for Large Model Inference},
  journal = {Electronics},
  year = {2026},
  volume = {15},
  number = {12},
  pages = {2567},
  doi = {10.3390/electronics15122567}
}
Kostenlos starten →