ConvXformer: Differentially Private Hybrid ConvNeXt-Transformer for Inertial Navigation
Omer Tariq, Muhammad Bilal, Muneeb Ul Hassan · 2025
Read ConvXformer: Differentially Private Hybrid ConvNeXt-Transformer for Inertial Navigation by Omer Tariq, Muhammad Bilal, Muneeb Ul Hassan (2025) 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-2510-19352.txt.
Description
Omer Tariq and four co-authors present ConvXformer. It is a model for inertial navigation with privacy built in. Motion sensors track position when satellite signals fail. Learned tracking beats older statistical methods in resolution. But training data maps people's movements. Such data is sensitive.
Standard privacy noise often ruins accuracy. Sensor signals change fast, which makes things worse. The hybrid design fuses convolutional blocks with transformer encoders. A new privacy method clips gradients in an adaptive way. It injects noise aligned with the gradient. Phones could run such models without exposing movement histories.
Catalog record
| CREATOR | Omer Tariq, Muhammad Bilal, Muneeb Ul Hassan |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 69 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2510-19352