Generative Adversarial Networks

Thursday 6 -7:45 pm // The course will provide you with hands-on experience about most recent machine learning models for generative adversarial networks. These networks are used to generate artificial photorealistic pictures and 3D objects

ANMELDEFRIST ABGELAUFEN

Für diesen Kurs kann man keine Teilnahmenbescheinigung erhalten.

Bei ausreichender Teilnahme und Erbringung eines der angebotenen Leistungsnachweise kannst Du ein Leistungszertifikat erwerben.

Kursinformation

KURSBESCHREIBUNG

This course is a hybrid course; you can participate either online via Zoom or in presence in Kiel (the number of places for presence participation is very limited).

Our courses are open to all and free of charge for students, pupils, trainees, founders, scientific staff, job seekers, refugees, and pensioners. For employees and self-employed persons, a fee of 200 euros is charged when acquiring a certificate of achievement. Employees and self-employed persons can apply for a scholarship via stipendium@opencampus.sh.

Your Background

You already have some knowledge about Machine Learning (ideally you followed one of our previous courses) and are interested in learning more about Generative Adversarial Networks. Particularly in this course you should be motivated to bring your own contribution – it's a great opportunity to learn and discuss latest discoveries in the field.

Course Content

This course is based on the Coursera Specialization on Generative Adversarial Networks (https://www.coursera.org/specializations/generative-adversarial-networks-gans).

The course will provide you with knowledge and experience about most recent and advanced machine learning models for Generative Adversarial Networks. You will be able to train and use neural network from the first week on.
You will then apply the newly acquired knowledge to implement your own machine learning project in a team.

How it works

The course will take place every Thursdays from 6 pm to 7:45 pm.

Before the first meeting there will be a semesterkick-off on the 08.04.2021.

During the week you will be expected to work through the assigned online course content, which will take you between 4 to 8 hours each week. Questions considering the course content and possible additional implications will then be discussed in the weekly online course meeting on Thursday. Towards the end of the semester you will then work in a team on your own machine learning project.
All needed software and online course content is free. For the practical assignments in the online course, however, it will be necessary to create a Google account.

The Formalities

In order to receive a certificate of attendance ("Leistungszertifikat") for this course, active participation is expected, and no more than two classes may be missed. The active participation is proven via the final presentation of your project by you and your team, and the delivery of a well documented project source code. The same conditions apply in order to receive ECTS.

In the online sessions it is necessary that you always provide your full name in Zoom so that your presence is registered on the EDU platform. No mere certificate of attendance will be issued for this course.

Further details may be given in the course.

ALLGEMEINE INFOS

Kursname:
Generative Adversarial Networks
Format:
Hybrid
ECTS:
5.0
Semester:
Sommersemester 2021
Anmeldefrist:
11.04.2021
Start:
15.04.2021
Ende:
17.06.2021
Zeiten:
18:00 - 19:45 Uhr

Kursleitung

Name:
Henrik Horst
E-Mail:
henrik.ho@gmx.de
Telefon:
0431/90894380
Start: 15.04.2021
Ende: 17.06.2021
General introduction about the course, the content and the projects.

Henrik Horst

See some real-world applications of GANs, learn about their fundamental components, and build your very own GAN using PyTorch!

Henrik Horst

Learn about different activation functions, batch normalization, and transposed convolutions to tune your GAN architecture and apply them to build an advanced DCGAN specifically for processing images!

Henrik Horst

Learn advanced techniques to reduce instances of GAN failure due to imbalances between the generator and discriminator! Implement a WGAN to mitigate unstable training and mode collapse using W-Loss and Lipschitz Continuity enforcement.

Henrik Horst

Understand how to effectively control your GAN, modify the features in a generated image, and build conditional GANs capable of generating examples from determined categories!

Henrik Horst

Understand the challenges of evaluating GANs, learn about the advantages and disadvantages of different GAN performance measures, and implement the Fréchet Inception Distance (FID) method using embeddings to assess the accuracy of GANs!

Henrik Horst

Learn the disadvantages of GANs when compared to other generative models, discover the pros/cons of these models—plus, learn about the many places where bias in machine learning can come from, why it’s important, and an approach to identify it in GANs!

Henrik Horst

Learn how StyleGAN improves upon previous models and implement the components and the techniques associated with StyleGAN, currently the most state-of-the-art GAN with powerful capabilities!

Henrik Horst

Learn different applications of GANs, understand the pros/cons of using them for data augmentation, and see how they can improve downstream AI models!

Henrik Horst

Presentation of the projects and follow-up discussion.

Henrik Horst