Artificial Intelligence and Machine Learning for Connected Systems

Code

USEEN6

Domaine

Informatique Télécoms Médias numériques Cybersécurité

Description

The course covers the following topics with half of the lessons as practical labs :

- refresh on statistics and network optimisation
- unsupervised machine learning
  - main algorithms, comparison, experimentation
  - time-constrained applications (traffic anomaly detection, etc)
  - memory-constrained applications (spatio-temporal mobility characterization, etc)
- supervised machine learning and applications
  - main algorithms, comparison, experimentation
  - time-and-energy-constrained application (IP traffic classification, etc)
  - time-and-memory-constrained applications (cyber attack classification, etc)

Finalité

The objective of the course is to study basics of machine learning and artificial intelligence algorithms used for network applications and IoT systems optimisation and acquire hands-on experience via experimental labs. The course will show how conventional ML/AI algorithms can be challenged in their performance and accuracy when running under constraints emerging in network and IoT systems environment, as for instance : execution time target, limited live and storage memory space, energy consumption and power limitations.