- Note regarding location: the department web page previously listed an incorrect location for this course.
The course meets on Monday, Wednesday 10:30AM - 11:45AM.
The correct location is East Hall, Room 015.
This course is offered in Fall 2016 as a full credit course.
The course will assume knowledge of material covered in comp136.
We will start with some lectures and shift to a seminar format.
Topics will include a more in depth coverage of variational inference, EP, and sampling methods, message passing in graphical models and its relation to variational inference, inference through optimization, inference as optimization tool for MDPs, black box inference methods, as well as more concrete models and algorithms for them (matrix factorization; topic models; etc). Additional topics to match students' interests.