Explore deep generative modeling in this comprehensive lecture from MIT's Introduction to Deep Learning course. Delve into latent variable models, autoencoders, and variational autoencoders, including the reparameterization trick and latent perturbation. Discover how VAEs can be used for debiasing, and learn about generative adversarial networks (GANs), their intuitions, and recent advances. Gain valuable insights into the importance and applications of deep generative modeling in machine learning and artificial intelligence.
Overview
Syllabus
- Introduction
- Why do we care?
- Latent variable models
- Autoencoders
- Variational autoencoders
- Reparameterization trick
- Latent pertubation
- Debiasing with VAEs
- Generative adversarial networks
- Intuitions behind GANs
- GANs: Recent advances
- Summary
Taught by
https://www.youtube.com/@AAmini/videos
