An Explainable Failure Prediction Framework for Neural Networks in Radio Access Networks
Khaleda Papry 1 , Francesco Spinnato 2,3 , Marco Fiore 4 , Mirco Nanni 3 , Israat Haque 1 · 2026
Read An Explainable Failure Prediction Framework for Neural Networks in Radio Access Networks by Khaleda Papry 1 , Francesco Spinnato 2,3 , Marco Fiore 4 , Mirco Nanni 3 , Israat Haque 1 (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-2602-13231.txt.
Description
Fifth-making networks promise high speed, low latency, and reliability, but millimetre-wave links fail easily under environmental effects, making uninterrupted service hard to guarantee. Khaleda Papry and four co-authors bridge an explainability gap: many radio link failure predictors work as black boxes, limiting operational trust. Their system pairs explainability-based feature pruning with model refinement and integrates with predictors such as graph neural network transformers and LSTM architectures.
Applied to radio link failure forecast, the system reveals weather data contributing minimally to forecasts in wide real-world datasets. That insight yields a leaner model with 50% fewer settings and improved F1 scores over the state-of-the-art answer. Network providers gain a way to evaluate and refine neural forecast models for interpretability, scalability, and scores, turning explainability from a reporting nicety into a design tool.
Catalog record
| CREATOR | Khaleda Papry 1 , Francesco Spinnato 2,3 , Marco Fiore 4 , Mirco Nanni 3 , Israat Haque 1 |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 51 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-13231