Deep Generative Modeling - MIT 6.S191 Lecture 4

This lecture from MIT's Introduction to Deep Learning 6.S191 series explores deep generative modeling, presented by Ava Amini. Dive into the fundamentals of generative models that can create new data samples resembling a training distribution. Learn about key architectures including autoencoders, variational autoencoders (VAEs), and generative adversarial networks (GANs), understanding how each approach tackles the challenge of generating realistic data. The 49-minute session is part of the 2025 edition of MIT's popular deep learning course. Access comprehensive materials including slides and lab exercises through the course website at introtodeeplearning.com, and follow @MITDeepLearning on social media for updates on new deep learning content from MIT.