Unlocking Interpretability for RF Sensing: A Complex-Valued White-Box Transformer Conference: ; 2025
Xie Zhang, Yina Wang, Chenshu Wu · 2025
Read Unlocking Interpretability for RF Sensing: A Complex-Valued White-Box Transformer Conference: ; 2025 by Xie Zhang, Yina Wang, Chenshu Wu (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-2507-21799.txt.
Description
Xie Zhang, Yina Wang, and Chenshu Wu build a wireless sensing model whose inner workings can be read like math, not guessed at like a black box. Radio signals can sense people through walls and darkness, but deep learning systems for the job are usually opaque, with parts chosen by trial and error.
Their model, RF-CRATE, derives every layer from a single compression principle, using numbers that carry both size and phase as radio signals do. A regularizing trick keeps its representations varied despite scarce labeled data, which lifted accuracy noticeably. Across five datasets covering activity, gait, gestures, pose, and breathing, it matched strong black-box rivals while staying interpretable. Gains were largest on the tasks where the physics of the signal matters most.
Catalog record
| CREATOR | Xie Zhang, Yina Wang, Chenshu Wu |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 96 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2507-21799