Lightweight Security-by-Design for TinyML Models in Constrained Embedded Environments
Haifa A. Alanazi, Khaznah R. Alshammari · 2026
Read Lightweight Security-by-Design for TinyML Models in Constrained Embedded Environments by Haifa A. Alanazi, Khaznah R. Alshammari (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/pmc-13611592.txt.
Description
Haifa Alanazi and Khaznah Alshammari proposed in 2026 a security design for tiny AI models that run on small devices. Such models sit inside sensors and wearables with very little memory and often no way to get a fix once deployed. Shrunk models also resist attacks less well than full-size ones, yet security is often bolted on after shrinking.
Their design has three layers. One scores models on accuracy, attack resistance, and size before the design is fixed. A second signs the model weights with a light cipher and checks them at start-up and on updates. A third watches live inputs for drift and raises defenses when needed. Tests on real hardware beat rival models.
Catalog record
| CREATOR | Haifa A. Alanazi, Khaznah R. Alshammari |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 45 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |