Introduction to Learning from Graph-Structured Data

Dates & Times: 29.06, 30.06 & 01.07  09:00-12:00 each day

Location: Seminar room 0.05, Hamburg Center for Translational Immunology (HCTI), Building N25, UKE

Instructor: Dr. Lucia Testa, Institute of Medical Systems Bioinformatics, UKE

Language: English

Prerequisites: Experience working in Python

Description: The aim of the course is to provide an introduction to  graph-based tools. Many real-world systems, including biological networks, molecular structures, and social interactions, are naturally represented as graphs. The workshop will introduce the fundamental concepts behind learning on graphs with a focus on graph theory, graph signal processing and graph neural networks. Participants will learn the mathematical foundations of graph-based learning and gain practical experience implementing models using PyTorch and PyTorch geometric. The course will combine theoretical explanations with hands-on tutorials. The workshop will be in presence and each participant should bring their own laptop.

Topics:

• Introduction to graph-structured data
• Fundamentals of graph signal processing
• Message passing and Graph Neural Networks
• Graph-level learning
• Implementing Graph Neural Networks
• Training and evaluating models on real datasets

Registration for this workshop will open soon.