The Radio-Frequency Transformer for Signal Separation
Egor Lifar, Semyon Savkin, Rachana Madhukara, Tejas Jayashankar, Yury Polyanskiy, Gregory W. Wornell · 2026
Read The Radio-Frequency Transformer for Signal Separation by Egor Lifar, Semyon Savkin, Rachana Madhukara, Tejas Jayashankar, Yury Polyanskiy, Gregory W. Wornell (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-2603-09201.txt.
Description
Egor Lifar and colleagues build a fully data-driven separator that pulls a signal of interest out of unknown non-Gaussian interference. They first learn a discrete tokenizer for the target signal, adapting a SoundStream design with extra transformer layers and finite-scalar quantization. A transformer trained with cross-entropy loss then beats mean-squared error training by a wide margin. On MIT RF Challenge mixtures the method cuts bit-error rates sharply and adapts to new interference with no side information.
Catalog record
| CREATOR | Egor Lifar, Semyon Savkin, Rachana Madhukara, Tejas Jayashankar, Yury Polyanskiy, Gregory W. Wornell |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 54393 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.09201