Logo ČVUT
Loading...
CZECH TECHNICAL UNIVERSITY IN PRAGUE
STUDY PLANS
2024/2025
Faculties | School elective courses | ECTS courses catalogue | Settings | Česká verze
Welcome page > Faculty of Information Technology > Department of Applied Mathematics > Graph Neural Networks
NOTICE: Study plans for the following academic year are available.

Graph Neural Networks

Display time-table
Code Completion Credits Range Language
NI-GNN Z,ZK 4 1P+1C Czech
Course guarantor:
Miroslav Čepek
Lecturer:
Miroslav Čepek
Tutor:
Miroslav Čepek
Supervisor:
Department of Applied Mathematics
Synopsis:

The course introduces students to advanced artificial intelligence techniques for working with graphs. Lectures will focus on the latest graph neural networks for creating vector representations of nodes, edges and entire graphs. The techniques discussed cover various types of graphs, including time-varying graphs. The last part of the course also covers graph generation and interpretability of graph neural networks. In the exercises, students will try out selected techniques and problems.

Requirements:

no entry requirements

Syllabus of lectures:

1) Introduction to the subject, motivation and definition of terms.

2) Representations based on the adjacency matrix and random walks through the graph.

3) Convolutional graph neural networks.

4) Representations of time-variable graphs.

5) Graph generation and representation using graph autoencoders.

6) Interpretability and applications in natural language processing and recommender systems.

Syllabus of tutorials:

1) Introduction to the StellarGraph library.

2) Vector representation of graphs.

3) Classification and clustering of nodes and graphs.

4) Graphs with time component.

5) Working on a semestral project.

6) Submission of the project and its presentation.

Study Objective:

The course introduces students to advanced artificial intelligence techniques for working with graphs. Lectures will focus on the latest graph neural networks for creating vector representations of nodes, edges and entire graphs. The techniques discussed cover various types of graphs, including time-varying graphs. The last part of the course also covers graph generation and interpretability of graph neural networks. In the exercises, students will try out selected techniques and problems.

Study materials:

Deep Learning; I. Goodfellow, Y. Bengio, A. Courville; MIT Press; 2016; ISBN 978-0262035613.

Introduction to Graph Neural Networks; Zhiyuan Liu, Jie Zhou; Morgan & Claypool Publishers; 2020; ISBN-13 978-1681737652

Graph Representation Learning (Synthesis Lectures on Artificial Intelligence and Machine Learning); William L. Hamilton; Morgan & Claypool Publishers; 2020; ISBN 978-1681739632

Heterogeneous Graph Representation Learning and Applications; Chuan Shi, Xiao Wang, Philip S. Yu; Springer; 2022; ISBN: 978-9811661655

Note:

blank

Further information:
https://courses.fit.cvut.cz/NI-GNN/
Time-table for winter semester 2024/2025:
Time-table is not available yet
Time-table for summer semester 2024/2025:
06:00–08:0008:00–10:0010:00–12:0012:00–14:0014:00–16:0016:00–18:0018:00–20:0020:00–22:0022:00–24:00
Mon
Tue
Wed
Thu
Fri
roomT9:303
Čepek M.
09:15–10:45
EVEN WEEK
(lecture parallel1)
Dejvice
roomT9:303
Čepek M.
09:15–10:45
ODD WEEK
(lecture parallel1
parallel nr.101)
Dejvice
The course is a part of the following study plans:
Data valid to 2025-03-12
Copyright © Jiří Kosek, 2010–2022 | Copyright © ČVUT, 2010–2022
Hosted by Výpočetní a informační centrum ČVUT
Send your questions and comments to HelpDesk ČVUT or to spravcekos@cvut.cz
For updated information see http://bilakniha.cvut.cz/en/predmet7023806.html