Bridging the Gap Between Simulated and Real Network Data Using Transfer Learning
Abstract
Machine Learning (ML)-based network models provide fast and accurate predictions for complex network behaviors but require substantial training data. Collecting such data from real networks is often costly and limited, especially for critical scenarios like failures. As a result, researchers commonly rely on simulated data, which reduces accuracy when models are deployed in real environments. We propose a hybrid approach leveraging transfer learning to combine simulated and real-world data. Using RouteNet-Fermi, we show that fine-tuning a pre-trained model with a small real dataset significantly improves performance. Our experiments with OMNeT++ and a custom testbed reduce the Mean Absolute Percentage Error (MAPE) in packet delay prediction by up to 88%. With just 10 real scenarios, MAPE drops by 37%, and with 50 scenarios, by 48%.
Cite
@article{arxiv.2510.00956,
title = {Bridging the Gap Between Simulated and Real Network Data Using Transfer Learning},
author = {Carlos Güemes-Palau and Miquel Ferriol-Galmés and Jordi Paillisse-Vilanova and Albert López-Brescó and Pere Barlet-Ros and Albert Cabellos-Aparicio},
journal= {arXiv preprint arXiv:2510.00956},
year = {2026}
}
Comments
This paper was submitted to IEEE NetSoft 2026. 7 Pages, 5 Figures