Katarzyna Wasielewska

José Camacho, Katarzyna Wasielewska, Pablo Espinosa et al.

Autonomous or self-driving networks are expected to provide a solution to the myriad of extremely demanding new applications with minimal human supervision. For this purpose, the community relies on the development of new Machine Learning (ML) models and techniques. %, like the celebrated Deep Learning (DL). However, ML can only be as good as the data it is fitted with, and data quality is an elusive concept difficult to assess. In this paper, we show that relatively minor modifications on a benchmark dataset (UGR'16, a flow-based real-traffic dataset for anomaly detection) cause significantly more impact on model performance than the specific ML technique considered. We also show that the measured model performance is uncertain, as a result of labelling inaccuracies. Our findings illustrate that the widely adopted approach of comparing a set of models in terms of performance results (e.g., in terms of accuracy or ROC curves) may lead to incorrect conclusions when done without a proper understanding of dataset biases and sensitivity. We contribute a methodology to interpret a model response that can be useful for this understanding.

José Camacho, Katarzyna Wasielewska, Rasmus Bro et al.

There is an increasing interest in the development of new data-driven models useful to assess the performance of communication networks. For many applications, like network monitoring and troubleshooting, a data model is of little use if it cannot be interpreted by a human operator. In this paper, we present an extension of the Multivariate Big Data Analysis (MBDA) methodology, a recently proposed interpretable data analysis tool. In this extension, we propose a solution to the automatic derivation of features, a cornerstone step for the application of MBDA when the amount of data is massive. The resulting network monitoring approach allows us to detect and diagnose disparate network anomalies, with a data-analysis workflow that combines the advantages of interpretable and interactive models with the power of parallel processing. We apply the extended MBDA to two case studies: UGR'16, a benchmark flow-based real-traffic dataset for anomaly detection, and Dartmouth'18, the longest and largest Wi-Fi trace known to date.