Lab Course: Graph Learning

Summer term 2019

Contact

Phone

work
+49 241 80 21706

Email

E-Mail

Content

In this practical course the participants will implement learning algorithms for graph data. This involves in particular topics such as graph kernels and graph neural networks which they will apply to different datasets.

Prerequisits

This pro seminar is only addressed to master students in the MSc Computer Science and MSc Data Science programmes.

No previous knowlege is required. Knowledge from the courses Algorithmic Foundations of Datascience, Algorithmic Learning Theory, Machine Learning are helpful, as well as knowledge of Python3 and Tensorflow.

Organization

The tasks will be worked upon by groups of two or three students.

Time and Place

Precise dates will be announced in RWTHonline before the semster starts.

The course will have a weekly meeting where results will be presented and possibly basics of graph learning will be presented.

Instructors

Martin Grohe

Assistants:
Hinrikus Wolf and Martin Ritzert

Requirements

Each participant of the seminar will be assigned a specific chapter from one of the books "Information and Coding Theory" by Jones & Jones and "A Concise Introduction to Data Compression" by Salomon. They are expected to give a talk of about 30 minutes about it and write a paper of about 5 pages summarising it.

The topics will be assigned at the first meeting of the seminar.

External Links