Cellular coverage quality estimation has been a critical task for self-organized networks. In real-world scenarios, deep-learning-powered coverage quality estimation methods cannot scale up to large areas due to little ground truth can be provided during network design & optimization. In addition they fall short in produce expressive embeddings to adequately capture the variations of the cells' configurations. To deal with this challenge, we formulate the task in a graph representation and so that we can apply state-of-the-art graph neural networks, that show exemplary performance. We propose a novel training framework that can both produce quality cell configuration embeddings for estimating multiple KPIs, while we show it is capable of generalising to large (area-wide) scenarios given very few labeled cells. We show that our framework yields comparable accuracy with models that have been trained using massively labeled samples.

翻译：蜂窝覆盖质量估计一直是自组织网络中的关键任务。在实际场景中，基于深度学习的覆盖质量估计方法由于在网络设计与优化期间可提供的真实标签极少，无法扩展到大规模区域。此外，这些方法难以生成具有充分表达能力的嵌入表示，以准确捕捉小区配置的变化。为应对这一挑战，我们将该任务建模为图表示形式，从而能够应用表现卓越的最新图神经网络。我们提出了一种新颖的训练框架，既能生成高质量的小区配置嵌入以估计多个关键性能指标，又能证明其在仅使用极少量标签小区的情况下，具备泛化到大规模（广域）场景的能力。研究表明，我们的框架在精度上与使用大量标签样本训练的模型相当。