Machine Learning for Networking
Passion in Action Course: Machine Learning for Networking
The activity is aimed at the presentation and experimentation of specific techniques of machine learning applied to the world of networks. The activity consists of two parts: a first theoretical part and a second more practical one.
The first part will discuss both the main techniques for collecting data and measurements from computer networks, and the main machine learning techniques with which such data can be analyzed and interpreted. In particular, the following topics will be presented:
- Active and passive network measurement methods
- Network data visualization techniques
- Machine learning techniques for data classification
- Applications of classification techniques to network data: traffic classification and anomaly detection
In the second part, the students will have to apply the previously illustrated techniques to a data set of cellular network measurements: the objective is to classify the cellular network users as satisfied or unsatisfied with the network service. Students organized in "teams", will have to adapt appropriately the machine learning algorithms studied during the first part, train them using the training data set and apply them to the validation data sets. The teams with the best accuracy in the classification of users will present their results to other participants.
The course is taught in collaboration with prof. Marcello Restelli and Francesco Trovò from the Computer Science section.
News
The new dates/classrooms are now available
Schedule
12/05/2020 - 17:30/19:30 - aula E.G.1: Introduction, active and passive measurements (link)
13/05/2020 - 17:30/19:30 - aula E.G.1: Data visualization (link)
14/05/2020 - 17:30/19:30 - aula E.G.6: Introduction Machine Learning and Supervised Learning (link)
15/05/2020 - 17:30/19:30 - aula E.G.1: Classification Methods (link)
19/05/2020 - 17:30 /19:30 - aula E.G.1: Exercise Session on Classification (link)
20/05/2020 - 17:30/19:30 - aula E.G.1: Bias/Variance Tradeoff (link)
21/05/2020 - 17:30/19:30 - aula E.G.6: Exercise Session on Bias/Variance Tradeoff (link)
22/05/2020 - 17:30/19:30 - aula E.G.1: Traffic classification (link)
28/05/2020 - 17:30/19:30 - aula E.G.3: WiFi Sniffing (link)
29/05/2020 - 17:30/19:30 - aula E.G.3: Final assignment and challenge (link)
Course Material
Slides
- Introduction on network measurements
- Introduction on machine learning and supervised learning
- Classification methods
- Exercise Session on Classification, datasets
- Bias/Variance Tradeoff, exercise
- Traffic classification
- WiFi Sniffing
- Final Challenge
Code examples
- Python 3 script for ping measurements (tested on Mac OS) - list of IP servers
- Machine learning examples
Other useful stuff
- Introduction to Statistical Learning (book by Hastie and Tibshirani)
- Elements of Statistical Learning (book by Hastie and Tibshirani)
- Machine Learning (online course by Andrew Ng)