Advanced Network Science (Autumn 2024)

Abstract

Networks are all around us: We are ourselves, as individuals, the units of a network of social relationships of different kinds; the Internet and the highway system can be modelled as networks embedded in space; networks can be also entities defined in an abstract space, such as networks of acquaintances or collaborations between individuals. This course aims at providing the computational tools to study these networks and form an advanced understanding of the current state of the art of network science. The final objective is to have the students master computational techniques to solve advanced network problems, to be able to contribute to the development of network analysis, and to appreciate the limitations and future developments of scientific papers dealing with network problems in real-world data.

 

Description

Network science is a thriving field of study. The reason of its popularity is its ability to represent very complex phenomena with very simple models (graphs). With network science one can analyze: global societal patterns, catastrophic breakdowns of distributed infrastructures, the metabolic pathways in humans at the basis of diseases, among many examples. The topics covered during the course focus on applying machine learning techniques to network science, and the relevant key network science topics required to adapt machine learning algorithms to this specific setting. The general topics are:

• Classical approaches to learning on graphs:
  • Node roles;
  • Node similarity and isomorphism.
• Machine learning on graphs:
  • Shallow Node Embeddings;
  • Deep Learning on Graphs;
  • Frequent Pattern Analysis.
• High-order network analysis:
  • Processes with memory;
  • Hypergraphs and simplicial complexes.
• Graph summarization.
• Statistics on complex networked spaces.

During the course, you will be reading articles applying machine learning techniques to network science, selected by the teacher. You will choose one as the main topic for your oral examination. 

Formal prerequisites

Intended learning outcomes

After the course, the student should be able to:

• Identify the relevant literature to follow the methodology of a network science research paper
• Reproduce the analysis of a network science research paper
• Select the most appropriate computational tool to address a specific question on a network
• Analyse real world social networks by applying advanced network analysis tools like (e.g. graph neural networks)
• Plan potential ways in which to customize an available tool to fit a new network feature or question

Learning activities

During the course, the instructors will provide readings instruction one week before each lecture: the students are supposed to read the material before class. Readings include chapters from the textbook and research papers. 

Each exercise session will involve dissecting the methodology behind a specific network analysis tool. Students are encouraged to hand-in their work and get feedback from teachers and TAs. 

Course literature

"The Atlas for the Aspiring Network Scientist" by M Coscia (course manager), available for free at: https://www.networkatlas.eu/ 

Additional material: pre-selected research papers, from the teacher (paper list still TBD), 

Student Activity Budget

Estimated distribution of learning activities for the typical student

• Preparation for lectures and exercises: 30%
• Lectures: 25%
• Exercises: 30%
• Exam with preparation: 15%

Ordinary exam

Exam type:
B: Oral exam, External (7-point scale)
Exam variation:
B1MH: Oral exam in group or individual, with time for preparation. Home.

reexam

Exam type:
B: Oral exam, External (7-point scale)
Exam variation:
B1MH: Oral exam in group or individual, with time for preparation. Home.

Time and date

Ordinary Exam Mon, 20 Jan 2025, 09:00 - 21:00
Ordinary Exam Tue, 21 Jan 2025, 09:00 - 21:00