Applications of Machine Learning methodologies for both supervised and unsupervised problems


Machine Learning (ML) is a branch of Artificial Intelligence (AI) that enables computers to gain knowledge from data. By digging into the data, a machine learning algorithm can get information about complex phenomena, catch correlations among variables, predict future instances of a sequence; it can also simplify complex data by collecting items into groups with “similar” properties and detect unexpected or anomalous patterns.  ML techniques can be supervised, whenever the learning process leverages on some data that are already labeled with information on their properties, or unsupervised, in which hidden patterns and structures are searched purely from data without any a-priori knowledge.

The applications are countless. In communications and networking, users behavior, their service demand, their mobility patterns, can be understood, predicted and classified  into similar groups, and then described through simplified models so as to improve service provisioning and optimize the use of resources. Anomalies and undesired events can be detected so as to engineer system protection mechanisms or speed up reactions to unexpected facts. The representation of contents like images and video, and their processing, can be improved in terms of quality, speed, efficiency.

ERC sectors 

  • PE6_7 Artificial intelligence, intelligent systems, natural language processing
  • PE6_11 Machine learning, statistical data processing and applications using signal processing (e.g. speech, image, video)
  • PE7_7 Signal processing


  • Machine learning
  • Deep neural network
  • Anomaly detection
  • Classification
  • Predictions
  • Predictive maintenance