Machine Learning for Graphs

In this course, we will introduce statistical methods to infer properties of large/complex networks. 

First, we will study how to define/learn standard properties of a network such as its diameter or connectivity. These properties are always useful to know, independently of their applications. 

Then, we will focus on two specific categories of networks, epidemic networks and engineering/telecom networks.In such networks, understanding general properties are not enough and specific problems have to be solved. 

In the case of epidemic network, we will study how to model the propagation of an epidemic and how we can estimate the parameter of our model. We will also show how to identify the source of an epidemic (patient zero).

Concerning engineering/telecom networks, we will teach you how to estimate the delay of communication on each link, in a telecom network. We will study how to detect the failure of a link.

In this course, we will provide answers to all these questions/problems using statistic and probability theory. It will be a combination of theory and practical assignments, where you will be asked to program different estimators on a real network.