清华主页 EN
导航菜单

Deep Generative Models

来源: 09-29

时间:Tues./Thur., 13:30-15:05,Oct.11-Dec.30,2022

地点:宁斋W11 Ning Zhai W11;Zoom ID: 330 595 3750

主讲人:Xie Pengtao

Target Audience:Undergraduate/Graduate students

Teaching Language:Chinese


Description:

This course introduces how to develop deep generative models (DGMs) by integrating probabilistic graphical models and deep learning to generate realistic data including images, texts, graphs, etc. Course contents include 1) basics of probabilistic graphical models, including Bayesian network and Markov random field; 2) posterior  inference methods, including message passing, variational inference, and Markov chain Monte Carlo sampling; 3) parameter learning and structure learning methods, including maximum likelihood estimation, expectation–maximization algorithm, and graphical LASSO; 4) deep generative models (DGMs), including variational auto-encoder, generative adversarial networks, normalizing flows, and  evaluation of DGMs; 5) applications of DGMs in  image generation, text generation, and graph generation.


Prerequisite:

Machine Learning, Probability and Statistics


Reference:

https://pengtaoxie.github.io/dgm.html


Bio:

Dr. Pengtao Xie's research interest is machine learning. He obtained his PhD in computer science from Carnegie Mellon University. He has published 50+ papers at top conferences and journals including ICML, NeurIPS, CVPR, ACL, etc. He was recognized as a top-5 finalist for the AMIA Doctoral Dissertation Award, Siebel Scholar, etc. He serves as area chairs for ICML, NeurIPS, etc.


返回顶部
相关文章
  • Deep Generative Models

    Description: This course introduces how to develop deep generative models (DGMs) by integrating probabilistic graphical models and deep learning to generate realistic data including images, texts, graphs, etc. Course contents include 1) basics of probabilistic graphical models, including Bayesian network and Markov random field; 2) posterior inference methods, including message passing, variat...

  • Deep image prior for inverse problems: acceleration and probabilistic treatment

    AbstractSince its first proposal in 2018, deep image prior has emerged as a very powerful unsupervised deep learning technique for solving inverse problems. The approach has demonstrated very encouraging empirical success in image denoising, deblurring, super-resolution etc. However, there are also several known drawbacks of the approach, notably high computational expense. In this talk, we des...