Evaluation of GANs

Objectives

• Differentiate across different evaluation metrics and their pros/cons.
• Justify the use of feature embeddings in GAN evaluation.
• Evaluate your GANs by implementing Fréchet Inception Distance (FID) and Inception Score.

GAN Disadvantages and Bias

Objectives

• Propose generative model alternatives to GANs and their pros/cons.
• Scrutinize bias in machine learning and examine its various sources.
• Describe an application of GANs that demonstrates bias.
• Explain several definitions of fairness in machine learning.

StyleGAN and Advancements

• Analyze key advancements of GANs.
• Build and compose the components of StyleGAN.
• Investigate the controllability, fidelity, and diversity of StyleGAN outputs.