GENERATIVE AI
Class overview
Generative AI are unsupervised machine learning methods that can create a wide variety of data, such as images, videos, audio, text, and 3D models. Deep generative models are the fundamental tools behind generative AI. They combine the generality of probabilistic reasoning with the scalability of deep learning. This course will study the probabilistic foundations and learning algorithms in generative AI, including variational autoencoders, generative adversarial networks, autoregressive models, normalizing flow, and diffusion models. This is a graduate-level course with an emphasis on mathematical principles as well as practical know-how. The course will be a combination of lectures, student presentations, and team projects.
Lecturer: Qi (Rose) Yu (roseyu@ucsd.edu)
TA: Jianke Yang (jiy065@ucsd.edu)
Time: 3:30 pm - 4:50 pm PT Tue, Thur
Location: EBU3B (CSE) 4140
Office Hours:
Jianke Yang | EBU3B B270A | Wed 9:00 - 10:00 am
Rose Yu | EBU3B 4216 | Mon 4:00 - 5:00 pm or by appointment
Syllabus
Week 1 (Sep 25) Introduction and deep learning recap
Week 2 (Oct 2) Generative AI background HW 1 release
Week 3 (Oct 9) Autogressive Models
Week 4 (Oct 16) Variational Autoencoder Project Proposal Due
Week 5 (Oct 23) Generative Adversarial networks HW 2 release
Week 6 (Oct 30) Normalizing Flow
Week 7 (Nov 6) Energy-Based Model Project Milestone Due
Week 8 (Nov 13) Diffusion Models HW3 release
Week 9 (Nov 20) Guest Lecture / Thanksgiving
Week 10 (Nov 27) Evaluation and Applications
Week 11 (Dec 4) Final project presentation Project Final Report Due
Lectures
Class Assessment
30 % homework (10% x 3)
50 % project
5 % proposal
15 % milestone report
20 % final report
10 % final presentation
20 % paper reading
Resources
Reading Materials
Other Tutorials
FAQ
Q: What are the pre-requisites?
Familiarity with statistical inference and deep learning.
Have taken at least an equivalent of CSE 250A and CSE 253.
Proficiency in Python.
Q: Can masters/undergraduates take this course?
This a graduate level audience.
Exceptional masters/undergraduates are welcome.
About me
My Chinese name is Qi Yu. That is also the instructor name in the registrar's office. I publish under the name Rose Yu. You can learn more about my research at my personal website.