Lecture notes for Graphical Models, COMPSCI688, taught by Justin Domke at UMass Amherst in Spring 2021.

These notes were primarily the effort of Nilesh Khade, Abhishek Ahuja, Abhinav Agrawal, Ke Xiao, and Justin Domke. These also include some figures based on assistants in previous versions of the course.

There are no doubt typos still remaining in these notes, and I'd still like to fix them, so please leave a comment if you find any. Comments should be available to anyone (though no promises on latency in making fixes!)

01 - Introduction - Feb 2, 2021

02 - Directed Models - Feb 4, 2021

03 - Representation in Directed Models - Feb 9, 2021

04 - Conditional Independency and Maximum Likelihood Learning - Feb 11, 2021

05 - Maximum Likelihood in Directed Models - Feb 16, 2021

06 - Directed and Undirected Models - Feb 18, 2021

07 - Hammersley-Clifford Theorem, Markov Random Fields - Feb 23, 2021

08 - Conditional Random Fields - Feb 25, 2021

09 - Message Passing - March 2, 2021

10 - Message Passing - March 4, 2021

11 - Exponential Family - March 9, 2021

12 - Exponential Family (Continued) - March 11, 2021

13 - Learning in Exponential Families - March 16, 2021

14 - Maximum Likelihood Efficiency - March 18, 2021

15 - Markov chain Monte Carlo - March 23, 2021

16 - Markov chain Monte Carlo (Continued) - March 25, 2021