Spring 2022 - Web 3:10-6:00 PM, Peking University

This course covers the fundamentals, research topics and applications of deep generative models.



Week 1 Introduction Lecture 1: Introduction
Lecture 2: Mathematic Foundation & Basic Concept
Lecture 3: Data Representation
Week 2 Autoregressive Models Lecture 4: Sequential Models - Recurrent Neural Networks
Lecture 5: Autoregressive Models 1
Lecture 6: Autoregressive Models 2
Week 3 Variational Autoencoders Lecture 7: From Autoencoder to VAE
Lecture 8: Variational Autoencoder

Lecture 9: VAE Variants
Week 4 Normalising Flow Models Lecture 10-11: Normalising Flow Background
Lecture 12: Normalising Flow Models
Week 5 Generative Adversarial Networks Lecture 13: Vanilla GAN
Lecture 14: Understanding GAN
Lecture 15: Selected GANs

Research & Application

Week 6 Evaluation of Generative Models Lecture 16-17: Sampling Quality
Lecture 17-18: Density Evaluation & Latent Representation
Week 7 Energy-based Models Lecture 19: Hopfield Network
Lecture 20: Boltzmann Machine
Lecture 21: Energy-based GANs
Week 8 Challenges of Generative Models Lecture 22: High-dimensional Data Generation
Lecture 23: Learning Large Encoder
Lecture 24: Other Challenges
Week 9 Applications of Generative Models Lecture 25: Image Synthesis, Translation and Manipulation
Lecture 26: X Learning
Lecture 27: Advanced Topics
Week 10 Discreteness in Generative Models Lecture 28-29: Discreteness in Generative Models - Discrete Sequence Generation.pdf
Lecture 29-30 Discreteness in Generative Models - Generating Graphs.pdf


Week 11 Paper Reading Score-based Generative Model
Graph Generation
Image Super-resolution with Deep Generative Models
Music Generation Models
Controled Text Generation
Graph Generative Model
Code Generation
Week 12 Paper Reading  
Week 13 Paper Reading  
Week 14 Group Projects  
Week 15 Group Projects  
Week 16 Group Projects  

Course Staff


For questions, please discuss on the Wechat group. You can also email Dr. Dong at hao.dong@pku.edu.cn.