CUSTOMER
Huawei
Starting from late 2016 we are conducting a research project focused on the classification of mobile and encrypted traffic. The project has two main objectives: (1) finding an efficient way to automate the generation of realistic annotated traffic traces (ground truth) from mobile apps; (2) designing novel approaches to classify in real time mobile apps traffic with high accuracy in real world scenarios (thousands of existing apps, frequent version updates, new apps released every day). The project is heavily focused on machine learning approaches for classification (supervised, semi-supervised, and non-supervised) and combination (hard and soft combiners), including those resulted very effective in other application domains (e.g., Deep Learning, Extreme Learning Machines).