Privacy Without Losing Place: A Paradigm for Private Retrieval in Spatial RAGs
Kennedy Edemacu, Mohammad Mahdi Shokri, Vinay M. Shashidhar, Jong Wook Kim · 2026
Read Privacy Without Losing Place: A Paradigm for Private Retrieval in Spatial RAGs by Kennedy Edemacu, Mohammad Mahdi Shokri, Vinay M. Shashidhar, Jong Wook Kim (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-05459.txt.
Description
Location-aware search tools need a user's position, and sharing it carries privacy risks. Kennedy Edemacu, Mohammad Mahdi Shokri, Vinay M. Shashidhar, and Jong Wook Kim propose PAS, short for Privacy Anchor Substitution. Instead of blurring the true location with noise, PAS describes position relative to an anchor, using a direction and a distance range.
On synthetic city data, attackers were left with a location error of roughly four hundred meters. The system kept over half of the baseline retrieval quality, and answers stayed fairly strong. The authors also found the privacy-utility balance behaves in a non-monotonic way.
Catalog record
| CREATOR | Kennedy Edemacu, Mohammad Mahdi Shokri, Vinay M. Shashidhar, Jong Wook Kim |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 62 KB |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.05459