Introduction to Generative Adversarial Networks (GANs)

Course details

Generative Adversarial Networks (GANs) are a groundbreaking innovation in deep learning, capable of generating realistic images, audio, and synthetic data. In this hands-on course, Gwendolyn Stripling—a machine learning and artificial intelligence and PhD—provides a solid foundation in the architecture and training of GANs. Learn how GANs work and get proactive practice, applying GANs to tasks such as synthetic data and image generation. This course will help you confidently step into the world of generative modeling and discover how these architectures are applied to real-world tasks.

Learning objectives
Explain the use cases, architecture, and function of GANs
Implement and train basic GANs using Python and Tensorflow/Pytorch libraries.
Apply GANs to real-world tasks, such as synthetic data and image generation.

Concepts

0. Introduction

01 - Transforming innovation with GANs
02 - What you should know

1. Introduction to GANs and Their Broad Impact

03 - Introduction to GANs and their broad impact
04 - Generative and discriminative models
05 - Use case - Synthetic customer reviews for product analysis

2. How GANs Learn to Create Generate Architecture

06 - The Generator
07 - The Discriminator
08 - How GANs learn

3. Building Your First GAN

09 - How to use the challenge exercise files
10 - Challenge - Build and evaluate a GAN
11 - Solution - Build and evaluate a GAN

4. GANs in the Real World

12 - GANs and multimodal LLMs for scalable content generation
13 - A collaborative pipeline

Conclusion

14 - Next Steps