SCU: An Efficient Machine Unlearning Scheme for Deep Learning Enabled Semantic Communications
Weiqi Wang, Zhiyi Tian, Chenhan Zhang, Shui Yu · 2025
Read SCU: An Efficient Machine Unlearning Scheme for Deep Learning Enabled Semantic Communications by Weiqi Wang, Zhiyi Tian, Chenhan Zhang, Shui Yu (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-2502-19785.txt.
Description
Weiqi Wang and three colleagues present a 2025 paper on teaching a communication system to forget. Semantic communication uses deep learning to send meaning rather than raw bits, training an encoder and decoder together. When users withdraw their private data, the trained pair must unlearn it, but existing unlearning methods assume a single supervised model.
Their scheme, SCU, attacks both halves at once. It scrambles the link between the learned representations and the samples to be erased. Then it repairs the damage by retraining, using the erased data as negative examples against the data that stays. Tests on three datasets showed effective erasure with the system's usefulness largely intact. The authors frame it as privacy hygiene for meaning-based networks.
Catalog record
| CREATOR | Weiqi Wang, Zhiyi Tian, Chenhan Zhang, Shui Yu |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 58 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2502-19785