Trans-RAG uses multi-stage query transformations to retrieve from mathematically isolated per-organization vector spaces, achieving 89.90° angular separation, 99.81% isolation, and only 3.5% nDCG@10 drop versus homomorphic encryption baselines.
In: Thirty- fifth Conference on Neural Information Processing Systems Datasets and Bench- marks Track (Round 2) (2021), https://openreview.net/forum?id=wCu6T5xFjeJ
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CR 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Trans-RAG: Query-Centric Vector Transformation for Secure Cross-Organizational Retrieval
Trans-RAG uses multi-stage query transformations to retrieve from mathematically isolated per-organization vector spaces, achieving 89.90° angular separation, 99.81% isolation, and only 3.5% nDCG@10 drop versus homomorphic encryption baselines.