Explore deep generative modeling in this comprehensive lecture from MIT's Introduction to Deep Learning course. Dive into the differences between supervised and unsupervised learning, understand the goals and applications of generative models, and learn about key concepts such as latent variables, autoencoders, and Variational Autoencoders (VAEs). Discover the intuition behind Generative Adversarial Networks (GANs), their training process, and their applications in image synthesis and paired translation. Gain insights into debiasing, outlier detection, and distribution transformations using these powerful deep learning techniques.
Overview
Syllabus
Intro
Supervised vs unsupervised learning Supervised Learning Unsupervised Learning
Generative modeling Goal Take as input training samples from some distribution and learn a model that represents that distribution
Why generative models? Debiasing
Why generative models? Outlier detection
What is a latent variable?
Autoencoders: background
Dimensionality of latent space reconstruction quality
Autoencoders for representation learning
Traditional autoencoders
VAEs: key difference with traditional autoencoder
VAE optimization
Intuition on regularization and the Normal prior
Reparametrizing the sampling layer
Why latent variable models? Debiasing
Generative Adversarial Networks (GANs)
Intuition behind GANS
Training GANs: loss function
GANs for image synthesis: latest results
Applications of paired translation
Paired translation: coloring from edges
Distribution transformations GANG
Taught by
https://www.youtube.com/@AAmini/videos
