Imran, AliBothe, Shruti2019-08-122019-08-122019-08https://hdl.handle.net/11244/321132Mobile cellular network operators spend nearly a quarter of their revenue on network management and maintenance. A significant portion of that budget, is spent on resolving faults diagnosed in the system that degrade or disrupt cellular services. Historically, the operations to detect, diagnose and resolve issues were carried out by human experts. However, with growing cell density, diversifying cell types and increased complexity, this approach is becoming less and less viable, both technically and financially. To cope with this problem, research on self-healing solutions has gained significant momentum in recent years. One of the most desirable features of the selfhealing paradigm is automated fault diagnosis. While several fault detection and diagnosis machine learning models have been proposed recently, these schemes have one common tenancy. They still rely on human expert contribution for fault diagnosis and prediction in one way or another. In this paper, we propose an AI-based fault diagnosis solution that offers a key step forward towards a completely automated self-healing system without requiring human expert input. The proposed solution leverages neuromorphic computing which uses RSRP map images of faults generated. We compare the performance of theproposedsolutionagainststateoftheartsolutioninliterature that mostly use Naive Bayes models, while considering seven different fault types. Results show that the neuromorphic model achieves high classification accuracy as compared to Random Forests classifier, Convolutional Neural Networks and Naive Bayes even with relatively small training data.Wireless Communication, 5G, Machine Learning, Self organizing networks